Black TiO2 nanotube arrays decorated with Ag nanoparticles for enhanced visible-light photocatalytic oxidation of salicylic acid

Article abstract of DOI:10.1016/j.jallcom.2018.10.248

Novel forms of black TiO₂ nanotubes-based photocatalysts for water purification were prepared. Two features were combined: decoration of TiO₂ nanotube arrays with Ag nanoparticles (sample TiO₂-NT's@Ag) and further hydrogenation of this material (TiO₂-NT's@Ag-HA). Obtained photocatalysts show high efficiency for degradation of salicylic acid, a typical water-borne pollutant. The photocatalysts considerably exceed the photocatalytic properties of TiO₂ nanotubes and commercial TiO₂ P25 taken as a reference for modeling of the photocatalytic process. The comparison of photocatalytic activities between novel photocatalyst was based on a numerical approach supported by the complex kinetic model. This model allowed a separate study of different contributions on overall degradation rate. The contributions include: salicylic acid photolysis, photocatalysis in UVB, UVA and in the visible part of applied simulated solar irradiation. The superior photocatalytic performance of the photocatalyst TiO₂-NT's@Ag-HA, particularly under visible irradiation, was explained by the combined effect of a local surface plasmon resonance (LSPR) due to Ag nanoparticles and creation of additional energy levels in band-gap of TiO₂ due to Ti³⁺ states at nanotube surfaces. The presence of Ag also positively influence charge separation of created electron-holes pairs. The synergy of several effects was quantified by a complex kinetic model through the factor of synergy, f_Syn. Stability testing indicated that the catalysts were stable for at least 20 h. The novel design of catalysts, attached on Ti foils, presents a solid base for the development of more efficient photocatalytic reactors for large-scale with a long-term activity.

Full text of DOI:10.1016/j.jallcom.2018.10.248

Accepted Manuscript
Black TiO nanotube arrays decorated with Ag nanoparticles for enhanced visible-
2
light photocatalytic oxidation of salicylic acid
Milivoj Plodinec, Ivana Grčić, Marc G. Willinger, Adnan Hammud, Xing Huang, Ivana
Panžić, Andreja Gajović
PII:
S0925-8388(18)33919-7
DOI:
https://doi.org/10.1016/j.jallcom.2018.10.248
JALCOM 48066
Reference:
To appear in: Journal of Alloys and Compounds
Received Date: 16 April 2018
Revised Date: 20 September 2018
Accepted Date: 20 October 2018
Please cite this article as: M. Plodinec, I. Grčić, M.G. Willinger, A. Hammud, X. Huang, I. Panžić,
A. Gajović, Black TiO nanotube arrays decorated with Ag nanoparticles for enhanced visible-
2
light photocatalytic oxidation of salicylic acid, Journal of Alloys and Compounds (2018), doi: https://
doi.org/10.1016/j.jallcom.2018.10.248.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT
Black TiO nanotube arrays decorated with Ag nanoparticles for enhanced
2
visible-light photocatalytic oxidation of salicylic acid
1
,3
2
3,4
3
3
Milivoj Plodinec , Ivana Grčić , Marc G. Willinger , Adnan Hammud , Xing Huang ,
1
1*
Ivana Panžić , Andreja Gajović
1
2
Ruđer Bošković Institute, Bijenička 54, HR-10000 Zagreb, Croatia
University of Zagreb, Faculty of Geotechnical Engineering, Hallerova aleja 7, HR-42000
Varaždin
Fritz-Haber-Institute der Max-Planck-Gesellschaft, Faradayweg 4-6, D-14195 Berlin,
Germany
3
4
Department of Colloid Chemistry, Max Planck Institute of Colloids and Interfaces, Am
Mühlenberg 1, 14476 Potsdam, Germany
ABSTRACT: Novel forms of black TiO nanotubes-based photocatalysts for water
2
purification were prepared. Two features were combined: decoration of TiO nanotube arrays
2
with Ag nanoparticles (sample TiO -NT’s@Ag) and further hydrogenation of this material
2
(
TiO -NT’s@Ag-HA). Obtained photocatalysts show high efficiency for degradation of
2
salicylic acid, a typical water-borne pollutant. The photocatalysts considerably exceed the
photocatalytic properties of TiO nanotubes and commercial TiO P25 taken as a reference
2
2
for modelling of the photocatalytic process. The comparison of photocatalytic activities
between novel photocatalyst was based on a numerical approach supported by the complex
kinetic model. This model allowed a separate study of different contributions on overall
degradation rate. The contributions include: salicylic acid photolysis, photocatalysis in UVB,
UVA and in the visible part of applied simulated solar irradiation. The superior photocatalytic
performance of the photocatalyst TiO -NT’s@Ag-HA, particularly under visible irradiation,
2
was explained by the combined effect of a local surface plasmon resonance (LSPR) due to
3
+
Ag nanoparticles and creation of additional energy levels in band-gap of TiO due to Ti
2
states at nanotube surfaces. The presence of Ag also positively influence charge separation of
created electron-holes pairs. The synergy of several effects was quantified by a complex
*
Corresponding author e-mail address: gajovic@irb.hr
1

ACCEPTED MANUSCRIPT
kinetic model through the factor of synergy, f . Stability testing indicated that the catalysts
Syn
were stable for at least 20 hours. The novel design of catalysts, attached on Ti foils, presents a
solid base for the development of more efficient photocatalytic reactors for large-scale with a
long-term activity.
KEYWORDS: black TiO nanotubes arrays, Ag decoration, salicylic acid photodegradation,
2
solar irradiation, kinetic model
2

ACCEPTED MANUSCRIPT
1
. INTRODUCTION
The TiO based materials are one of the most studied materials in the last 40 years,
2
with the applications in various fields such as: solar cells [1], photocatalysis [2], water
splitting [3], batteries [4], for bone and dental implants due to biocompatibility [5], gas
sensing due to change in electrical and optical properties [6]. On the other hand, TiO2
nanotubes are widely investigated from its beginnings in 2001, when Grimes and co-workers
for the first time synthesized TiO nanotube arrays by electrochemical oxidation of Ti-foil
2
[
7]. It is a relatively simple process and it is possible to apply the same procedure on the large
scale to grow uniform array of TiO nanotubes. Advantages of obtained TiO nanotubes are
2
2
temperature and chemical stability, relatively low cost of production, film flexibility, and
high specific surface area [8]. The anodization process enables to precisely control the
morphological parameters of thin film growth: film thickness, nanotubes diameter, wall
thickness, surface roughness [8]. Those parameters can be controlled by variation of: applied
DC voltage, type of electrolyte, the concentration of water and fluoride ions, reaction time
[
8]. The anodization parameters also influence the adhesion of the TiO nanotube arrays to Ti
2
substrate, so the adhesion can be enhanced by two-step anodization process described by Yu
et al. [9]. Advantages of TiO nanotube arrays compared to other TiO morphologies are a
2
2
high specific area (many reaction centers at the nanotube surface), good carrier mobility due
to the wire like morphology, and environmental stability [10]. The photocatalytic activity of
pure TiO nanotubes is limited due to the relatively large band gap of 3.2 eV for anatase and
2
3
.02 eV for rutile (bulk materials) [11]. In order to increase activity, it is essentially to modify
the surface; to dope or to decorate the nanotubes with the aim to tune band gap in the desired
range of solar irradiation absorption (visible and near infrared, NIR) [12].
3

ACCEPTED MANUSCRIPT
In the last 30 years there were many different approaches and attempts to increase photo
catalytic activity of TiO by doping with non-metal and metal atoms [13-23], depositing or
2
decorating the surface of TiO nanostructure (nanoparticles, nanowire, nanotubes, nanobelts)
2
by noble metal nanoparticles (Pt, Pd, Au, Ag) or metal oxide nanoparticles (WO , Cu O,
3
2
CeO , Fe O ) [23-33]. The electron trapping in the junction between noble metal
2
2
3
nanoparticles and TiO is determined by the work function of the metal, which is usually
2
larger than in TiO . The electron-hole pair can be formed by photon absorption at the surface
2
of TiO , formed electron would be transferred from conduction band (CB) of TiO to metal
2
2
nanoparticles. Schottky barrier, which is formed at the junction, leads to efficient charge
separation due to the existence of an internal electric field between TiO and metal
2
nanoparticles and acts as a trap for the generated charge carriers.
The photocatalytic materials decorated by noble metals are classified as plasmonic
photocatalyst [34]. The noble metal nanoparticles can be uniformly deposited (Au, Ag, Pd,
Pt) by one of the chemical methods with the particles size in the range of a few nanometers
up to several 100 nm. The aim of deposit nanoparticles at the surface of the nanotubes is to
increase UV and visible light absorption, to ensure better charge carrier separation, an
increase of local electric field and localized surface plasmon resonance (LSPR). By the
variation of the nanoparticles size, it is possible to tune the wavelength of absorption
maximum [34]. The embedding of the noble metal nanoparticles at the surface of the
semiconductor provide several beneficial effects. In the plasmonic photocatalysis, there are
two distinct features which contribute to better performances of the plasmonic photocatalysts
in comparison to conventional ones. Those are: (i) Schottky junction and (ii) localized surface
plasmon resonance (LSPR). There is forced electron-hole separation close to the interface
and fast lane charge transfer due to the first feature. While LSPR induced several more
effects: visible light response, enhanced UV-vis absorption, reduced electron-hole diffusion
4

ACCEPTED MANUSCRIPT
length, enhanced local electric field, LSPR electron-hole generation, local heating effects and
molecule polarization effect [34]. Yu et al. [35] prepared a new metal–semiconductor
nanocomposite plasmonic photocatalyst by depositing AgCl nanoparticles into the self-
+
organized TiO nanotubes, followed by partial reduction of Ag ions in the surface region of
2
AgCl particles. Such photocatalyst showed a high visible-light photocatalytic activity for the
photocatalytic degradation of methyl orange. Different authors studied the influence of the
particle size of noble-metal elements (Ag, Au) that decorate the TiO nanostructures on the
2
intensity and red shift of the absorbance in visible region [36, 37].
The ‘’black’’ TiO nanostructures obtained by annealing in the reductive atmospheres (H ,
2
2
Ar/H , CO, NO) has drawn a lot of attention recently. Reduced TiO nanostructures have
2
2
better properties for applications in DSSC cells, water splitting and photocatalysis compared
to non-reduced TiO [38, 39]. One of the reasons is narrowed band gap (< 3 eV), which
2
allows the photocatalytic activity of TiO in the visible range of solar irradiation. The
2
narrower band gap could result in a higher efficiency in the photocatalytic reactions [40].
3
+
Recent studies gave an explanation in terms of creation of Ti states during thermal
treatment of TiO nanotubes in the reductive atmosphere [41, 42]. High-temperature
2
annealing in the inert oxygen-free atmosphere, in a vacuum or reductive atmosphere lead to
3
+
the release of O molecules and formation of Ti ions, usually at the nanotubes surfaces.
2
Reduced nanotubes showed absorption in the visible part of solar spectra and better
conductivity. These effects can be explained through the appearance of the high density of
localized donor states which are created in forbidden gap of TiO due to the formation of
2
3
+
Ti , which actually represents electron donor state closely to the bottom of the conduction
band of TiO . The high-pressure H treatment of TiO nanotubes lead to anatase-like “black
2
2
2
titania”, showing a high open-circuit photocatalytic hydrogen production rate without the
3
+
presence of a co-catalyst. Through these processes, it is possible to create the most of 1% Ti
5

ACCEPTED MANUSCRIPT
4
+
from Ti , before crystal structure collapsed. This process is reversible till some point,
3
+
because exposure to air or O atmosphere leads to oxidation. Similar as Ti , the oxygen
2
vacancies also increase absorption in the visible part of solar spectra. Chen and co-workers
firstly reported on the synthesis of “black” TiO nanoparticles by annealing in the reductive
2
atmosphere (H atmosphere at 200 � C for 5 days and pressure 20 Bar) [38]. The
2
nanoparticles had a core-shell structure with crystalline TiO inner part of nanoparticles and
2
outer amorphous shell. Naldoni et al. synthesized TiO nanocomposites with core-shell
2
structure, with crystalline core and amorphous 2 nm thick shell [43]. They explained
3
+
narrowing of the bandgap by the existence of defects such as Ti and oxygen vacancies
which create energy states inside the bandgap. These states allowed the absorption and
intraband transition in the energy region smaller than 3.2 eV. Tao et al. applied a different
approach to obtain narrowed bandgap of 2.1 eV. They obtained TiO phase forms on the
2
surface of rutile TiO (011) by oxidation of bulk titanium interstitials [44], which resulted in
2
3
+
the similar photocatalytic activity due to TiO self-doping with Ti during annealing process
2
in a reductive atmosphere of CO or NO [45]. Wang et al. proposed that H-doped amorphous
shell plays the same role as Ag or Pt loading on TiO nanocrystals, which induces the
2
localized surface plasmon resonance and black coloration, resulting in enhanced
photocatalytic activity [46]. Wu et al. [47] applied the electrochemical hydrogenation doping
for the introduction of interstitial hydrogen ions and oxygen vacancies. Similar process was
4
+
3+
used by Zhao et al. [48] to reduce Ti to Ti , that is approved by XPS measurements. In our
previous work [49], powder of titanate (H Ti O ) nanotubes decorated by Ag was studied for
2
3
7
photocatalytic properties. We have shown that annealing of the protonated type of titanate
nanotubes in Ar/H induce a transition to TiO anatase phase nanoparticles can be used as
2
2
efficient powder catalyst. Moreover, H Ti O nanotubes decorated with Ag nanoparticles
2
3
7
6

ACCEPTED MANUSCRIPT
after annealing in the reductive atmosphere (Ar/H ) had two times higher photocatalytic
2
activity than those which were annealed in Ar/H but not decorated with Ag
2
With the aim to prepare the efficient easy-handling photocatalyst for purification of waste
water from the industry, we studied the modification and decoration of titanium oxide
nanotube arrays (TiO -NT’s) that are completely attached on titanium foil. In that manner,
2
there are no nanoparticles left in purified water as in the case of the use of suspended powder
catalyst (e.g. P25 TiO ). To increase the efficiency of the TiO -NT’s photocatalyst caused by
2
2
LSPR, titanium oxide nanotubes arrays were decorated by Ag nanoparticles (sample TiO2-
NT’s@Ag) and subsequently reduced by annealing in hydrogen (TiO2-NT’s@Ag-HA) with
the aim to narrow band gap of titania. The photocatalysts were evaluated by photodegradation
of salicylic acid (SA) in aqueous solution. SA was chosen as a model pollutant due to its
widespread use in dyestuff, resins and aspirin production, which results in its release to the
environment through various waste streams. SA was listed as a pollutant in precipitation,
-1
surface waters (approx. 0.1 µg L ) and as a constituent of humic material in drinking waters.
It may be released to the aquatic environment in wastewater discharges from industry and
even sewage treatment facilities [50].
The TiO -NT’s arrays were modified by Ag using similar procedures as in our previous
2
research [49]. In this work it was shown that TiO nanotubes formed by anodization process
2
on Ti foil decorated with Ag nanoparticles after annealing in the reductive atmosphere
(
TiO @Ag-HA) have the best overall photocatalytic performance, even 1.53 times larger than
2
commercial P25. The increase in the efficiency is explained by the combined effect of
localized surface plasmon resonance (LSPR) of the silver nanoparticles and the increased
absorption of visible light and NIR due to the reduction of TiO in hydrogen and formation of
2
3
+
defects such as Ti and oxygen vacancies. These defects create energy states inside the
7

ACCEPTED MANUSCRIPT
bandgap that allow an interband transition in the energy region smaller than 3.2 eV that can
be achieved by visible light.
The kinetic study presented in this work resulted in a complex kinetic model which
allowed a separate study of the photochemical degradation of salicylic acid and photon
absorption effects in UVB, UVA and visible part of applied simulated solar irradiation. These
discrete photon absorption study supported detailed analysis and quantification of
fundamental photocatalytic phenomena. Moreover, the absorption in UVB and UVA part of
applied irradiation was described with relative coefficients with TiO P25 as a reference.
2
Contribution from the absorption terms was accounted in the overall degradation rate
equations, along with the factor of synergy, f . Introduction of f . in the model, resulted in
Syn
Syn
quantification of the influence of underlying phenomena on the activity, other than pure
3
+
absorption of incoming irradiation, e.g., surface defects as Ti and LSPR.
2
. MATERIALS AND METHODS
.1 Materials
2
The following materials were used: titanium foil (10x10 cm, purity 99.5 %, Sigma
Aldrich), Ethanol (Fluka, 98%), deionized water, AgNO (Kemika, purity 99.8%), gas
3
mixture Ar/H (95% Argon,5% hydrogen, Messer, purity 5.0). Salicylic acid (SA) p.a. the
2
grade was obtained as a free-of-charge sample from Pliva, Croatia and used without further
-1
purification. A model solution of 0.2 mmol L of SA was prepared (pH 4). The degradation
of SA in the samples was evaluated in terms of SA concentration decrease and mineralization
degree. The concentration of SA was determined using HPLC, Shimadzu, with SUPELCO
C18 column, length 250 mm, internal diameter 4.6 mm and UV (diode array) detection
8

ACCEPTED MANUSCRIPT
(DAD) at 303 nm. The analyses were carried out by binary flow elution with mobile phase
−
1
solvents A and B at a flow rate of 0.6 and 0.4 mL min , respectively. Solvent A consisted of
0
.5:99.5 (v/v) methanol: 0.1 M phosphoric acid and solvent B was 100% methanol. This
method allows simultaneous determination of SA and its degradation by-products: 2,5-
dihydroxybenzoic acid (gentinsic acid), 2,3-dihydroxybenzoic acid, 3,4,5-trihydroxybenzoic
acid (gallic acid), hydroquinone, catechol and resorcinol using multi-wavelength DAD. The
sum of their concentration expressed through the C atom balance was used to evaluate the
conversion of aromatics to lumped short-chain carboxylic acids. The mineralization degree
was determined by measuring the total organic content of model wastewater by Total Organic
Carbon analyzer, TOC-VCPN, Shimadzu, Japan.
2
.2 Synthesis
The TiO nanotube arrays were obtained by anodizing a titanium metal foil (purity 99.7%,
2
size 1.5x1.5 cm and thickness 0.25 mm). The anodization was carried out at room
temperature in a conventional two-electrode cell using a direct current (DC) power supply A
Pt foil was used as the counter electrode and a Ti-foil served as the anode. The electrolyte
was ethylene glycol containing 0.3 wt.% NH F and 2 vol.% distilled water. The self-
4
organized and well-aligned TiO nanotube arrays were fabricated at 60 V for 3 h. As a result
2
of the anodization process, the TiO nanotubes were etched of in the titanium foil. The details
2
of chemical reactions and phenomena acting during formation of a nanotubes by anodization
are described in our previous work [51]. Following anodization, the specimens were rinsed in
deionized water and ethanol, and dried in a cool air stream. The anodized specimens were,
then, thermally treated in a tube furnace at 500 � C (heating/cooling rate of 2 � C/min) for 2
h. The anodization parameters stated above were found by optimization of the anodization
9

ACCEPTED MANUSCRIPT
process to obtain samples with TiO layers that were attached to the Ti surface strong enough
2
to resist all the future treatment.
The photo-reduction process was used for decoration of TiO nanotubes with Ag
2
nanoparticles, similar to one explained in our previous work [46]. Precisely, the Ti-foils with
thin film TiO nanotubes arrays (TiO NT’s) were immersed in 0.05 M solution of AgNO ;
2
2
3
the solution was continuously stirred under UV light for 3h at 80 � C. After the decoration of
TiO nanotubes with Ag nanoparticles (TiO NT’s@Ag), the foil was rinsed with ethanol and
2
2
deionized water and dried for 1 h at 80� C in air. TiO NT’s and TiO NT’s@Ag samples
2
2
were annealed in a reductive atmosphere (Ar/H ) for 3h at 500 � C, temperature rate was 5
2
°
C/min. Obtained samples were assigned as TiO NT’s-HA and TiO NT’s@Ag-HA.
2 2
2
.3 Characterization
The Raman spectroscopy (RS) measurements were performed using Horiba, Jobin–Yvon
T64000 spectrometer with argon-ion laser Coherent, Innova 400 operating at 514.5 nm for
the excitation. The Raman spectra were collected in micro- Raman mode with a multi-
channel CCD detector. A laser power of 20 mW at the sample and an objective with a 50
magnification were used.
The morphology, surface and chemical composition of modified samples were studied by
the scanning electron microscopy (SEM) using a Hitachi S-4800 with a cold FEG (Field
Emission Gun). The images were taken in secondary electron (SE) mode with an acceleration
voltage of 15 kV. The energy dispersive spectroscopy (EDS) analysis was also done at 15
kV.
For the transmission electron microscopy (TEM) analysis modified samples were prepared
in the cross-section by Gatan precision ion polishing system (PIPS) and ion beam etching
system. The TEM, high-resolution TEM (HRTEM) and high angle annular dark field
10

ACCEPTED MANUSCRIPT
scanning transmission electron microscopy (HAADF STEM) imaging were performed on the
TEM Philips CM200F and FEI TITAN 80-300 kV.
For the EELS data acquisition, double Cs aberration corrected (TEM and STEM) JEOL-
ARM 200CF equipped with GIF Quantum Energy Filters spectrometer for the high resolution
and DualEELS spectra acquisition was used. Measurements were performed under following
conditions: dispersion 0.1 eV/channel and integration time for each pixel on EELS maps of 2
seconds. The pixel size was 1.5 nm. Afterwards, the collected spectra were corrected for dark
current and channel-to-channel gain variation. The samples of EELS measurements were
prepared in lamela’s form by the FEI-Helios Dual-beam system.
The UV–vis-NiR diffuse-reflectance spectroscopy was utilized for the optical reflectance
characterization and band-gap calculation of the pure and Ag-decorated synthesized TiO2
nanotubes before and after annealing in the hydrogen atmosphere. The optical absorption
spectra were measured with a UV–vis-NiR Perkin–Elmer Lambda 650 high-performance
spectrophotometer a wavelength range from 300 nm to 900 nm, with a step size 1nm. The
samples were mounted by the scotch tape at the sample holder and analyzed in reflective
mode with the aperture 0.5 x 1.0 cm.
2
.4 Photocatalytic activity
Photocatalytic experiments were performed in a flow cell (4.5 cm in diameter, 2 cm in
height) working in total recirculation using a peristaltic pump (Vtotal = 40 mL, flow rate Q =
-1
6
7 mL min ). Plates with the photocatalytic film were put on the cell bottom, whereby
directly irradiated [52], so the catalysts were positioned perpendicular to the light irradiation.
-3
Prior to experiments, SA solutions (C0, SA = 0.5, 0.2, 0.1 mmol dm ) were recirculated in the
dark over the plates to achieve the sorption equilibrium. The sources of irradiation were: (i)
full-spectrum compact fluorescent bulb simulating solar spectra with high UVB; high colour
11

ACCEPTED MANUSCRIPT
rendering Ra98/Class 1A with colour temperature of 6500K (Exo Terra, 20W), and (ii)
incandescent bulb emitting only visible light (Osram Daylight, 100W). The UVB and UVA
intensities were measured by UVP UVX radiometer, fitted with the corresponding UVB and
UVA sensors. For the incandescent bulb, both UVA and UVB intensities were measured as
-2
0
.0 mW cm , indicating a complete absence of UV radiation in the spectrum of this bulb. For
the full-spectrum bulb, given intensities on the cell top (at free water surface) were: IUVB,0
=
-2
-2
2
.92 mW cm and I
= 3.43 mW cm . Intensities at film surface which were accounted
UVA,0
-2
for the incident photon flux at photocatalysts surface were: IUVB,f = 0.18 mW cm and I
UVA,f
-2
=
2.65 mW cm . Note that is much lower than IUVB,0, due to the absorption of light by SA.
The IUVB,f was obtained from radiometric readings from the bottom of the cell filled with the
SA solution. Experiments were performed in triplicates to discard possible experimental error
and to check the reusability of the photocatalysts. With the aim to study the stability of the
catalysts additional ten consecutive cycles of photocatalytic experiments were done using
full-spectrum compact fluorescent bulb simulating solar spectra.
3
. NUMERICAL APPROACH FOR QUANTIFICATION OF PHOTOCATALYTIC
ACTIVITY
The novelty presented in this work is the quantification of the photocatalytic activity of
prepared materials, compared to a reference, TiO P25. The latter was used as a reference
2
since it is the most common commercial photocatalysts with an appropriate activity. The the
main hypothesis in this study was the achievement of advantageous photocatalytic properties
of TiO NT’s@Ag and TiO NT’s@Ag-HA compared to anatase film. Tentatively, the
2
2
photocatalytic properties of these films would also exceed the properties of the efficient
12

ACCEPTED MANUSCRIPT
commercial TiO , i.e., TiO P25 in the case when the same was prepared in the form of thin
2
2
film. Having in mind the specific form of novel photocatalyst and inability to produce the
film having only pure TiO P25 with similar morphological characteristics, the comparative
2
quantification was done using the numerical method according to the algorithm presented as
follows.
The detailed kinetic model was developed to include the contribution from direct
photolysis as well as the contribution from UVB, UVA and visible light to photocatalytic
degradation of SA.
-3 -1
The total rate of SA degradation (rSA, T, kmol m s ) can be described as follows:
UVB
ph
UVA
vis
r
= r + r
+ r
+ r
SA,T
p
ph
ph
1
44424443
r
- photocatalysis
ph
(1)
In Eq. (1), r is the rate of direct photolysis of SA under UVB irradiation, r is the rate of
p
ph
photocatalytic degradation where superscripts UVB, UVA and vis denominate contributing
term. Note that pseudo-monochromatic irradiation (310 and 365 nm) was assumed for UVB
and UVA part of irradiation in further calculations. This simplification can be found in
literature as well [53].
To define the rate of photolysis, near-surface light absorption by SA was assumed. First-order
-1
photolysis rate constant k (s ) was given, Eq. (2) [54];
p
A
k
(
λ
)
= 2.3Φ
(
λ
)
ε λ
(
)
W
(
λ
)V
l
p
SA
SA
(2)
where ΦSA is the photochemical quantum yield for SA photolysis, εSA is the molar extinction
-1
-1
-2 -1
coefficient (L mol cm ), W is the incident light intensity (Einstein cm s ), A, V and l are
the surface of the reaction cell, total volume and solution depth, respectively. The λ stands for
the wavelength of concern, which in this case is exactly 310 nm, taking into account the
assumed monochromatic irradiation over UVB region. The photochemical quantum yield can
13

ACCEPTED MANUSCRIPT
be calculated using the observed rate of SA photolysis. The latter will be reported in the
Results and Discussion section.
k
p, obs
Φ
=
SA, 310nm
2.3ε
W
(
A
/
)
V l
SA, 310 nm UVB
(3)
The molar extinction coefficient was obtained from spectrophotometric analysis of SA
solutions of different concentration (data not shown). The εSA, 310 nm was found to be 2710 L
-1
-1
mol cm . The incident light intensity was obtained from Eq. (4).
I
UVB,0
W
=
UVB
hc
N
A
λ
(4)
-2
The IUVB, 0 (Wm ) is the irradiation intensity at the free water surface determined by
radiometric measurements, N , h and c represents Avogadro’s number, Planck constant and
A
-5
speed of light, respectively. Resultant incident light intensity at 310 nm was 7.562 × 10
-
2 -1
Einstein cm s .
-8
Finally, quantum yield was ΦSA, 310 nm = 8.59 × 10 . Note that calculation was made per unit
of the reaction cell, meaning that reported quantum yield can be used for calculation of
photolytic rates in other reactor configurations.
The rate of photocatalytic reaction can be written as follows:



m1
m2
m3



vis
r
= −k
µ
I
+
µ
I
+ P
C
SA
ph
SA, int
(
310nm UVB, f
)
(
365nm UVA, f
)
(
)
ph
(5)
-1
-0.5
In Eq. (5). kSA, int (m s W ) represent the intrinsic reaction rate constant for SA degradation
vis
over irradiated photocatalytic films. Contributing terms: ꢀ3
I
, ꢀ
I
and Pph
10nm UVB,f
365nm UVA,f
represent the rate of photon absorption in UVB, UVA and visible part of simulated solar
spectra, respectively, which are proportional to the rate of reaction [55,56]. The exponents
m , m and m represent the order of the reaction with respect to the rate of photon absorption
1
2
3
14

ACCEPTED MANUSCRIPT
in related part of solar spectra. The half-order dependence in well-irradiated system was
assumed (m = m = m = m = 0.5).
1
2
3
To estimate the relative precedence of studied films compared to the reference, the rate
equation was modified to Eq. (6). It needs to be noted that well-mixed, lump reaction system
with a uniform distribution of incident irradiation on the film surface was considered.
m
vis
r
= −k
µ
'
µ
I
+
µ
'
µ
I
+ P
C
SA
ph
SA, int
(
app,310nm P25,310nm UVB, f
app,365nm P25,365nm UVA, f
)
ph
(
6)
Herein, ꢀ’app,310nm and ꢀ’app,365nm are the apparent relative absorption coefficient of the film
=ꢀ / ꢀ ). Those numbers simply suggest how many times the studied film absorb more
(
film
P25
irradiation than the referent TiO P25. The ꢀ
and ꢀP25, 365 nm are referent absorption
2
P25, 310 nm
coefficients calculated from the specific absorption coefficient for TiO and its mass per
2
square meter of the film (= κ*m /A ). The specific absorption coefficients (κ*) for TiO P25
cat
f
2
can be found in the literature [57-59]. Used values are summarized in Table 1. Mass of the
-5
-2
photocatalytic film per area of the film was estimated at 5.66 × 10 kg m . Finally, the term
vis
-2
Pph (in W m ) represent the overall rate of photon absorption in the visible part of
simulated solar spectra. This term was determined experimentally from the experiments
where only the visible light was used, Eq. (7).
2



vis
SA,obs



k
vis
P
=
ph
k
SA, int




(7)
vis
SA, obs
Note that k
are observed pseudo-first order reaction rate constants for photocatalytic
degradation of SA over TiO NT’s, TiO NT’s@Ag and TiO NT’s@Ag-HA under visible
2
2
2
light (presented in the following section), while the intrinsic reaction rate constant was
-7
-1
-0.5
1.5
reported in our previous work [57], where it was expressed as 4 × 10 min W m . In this
-8
-1
-0.5
study, the surface rates are used and the kSA, int was recalculated to 1.4 × 10 s W m.
15

ACCEPTED MANUSCRIPT
vis
Resultant photon absorption in the visible part of the spectra, Pph , values are shown in
Table 1. The presented modelling approach is however incomplete. By expressing the total
photocatalytic rates with Eq. (6), one can assume that only the absorption of light in a visible
part of solar spectra and the enhanced absorption in UVB and UVA were responsible for the
enhancement of the overall SA degradation rates. The absurd values of ꢀ’app,310nm and
ꢀ’app,365nm, approaching several thousands and millions (Table 1), suggest the incorrect
assumptions. Having in mind previously mentioned the influence of the combined effect of
3
+
local surface plasmon resonance of Ag and creation of amorphous layer with Ti and oxygen
vacancies states at nanotubes surfaces, the rate can be written as follows; introducing the
overall factor of synergy (f ) into equation (Eq. (8)).
Syn

m



vis
r
= − k
µ
"
µ
I
+
µ
"
µ
I
+ P
f C
Syn SA
ph

SA, int
(
app,310nm P25,310nm UVB, f
app,365nm P25,365nm UVA, f
)
ph

(
8)
Hereby, ꢀ”app,310nm and ꢀ”app,365nm are the apparent relative absorption coefficient of the film
=ꢀ / ꢀ ). For simplification purpose, the apparent relative absorption coefficient for TiO
(
film
P25
2
NT’s was assumed to be 1. This assumption was justified by the similarities in the optical
properties between TiO NT’s and TiO P25 in the range from 300 – 400 nm (Fig. 7).
2
2
Namely, for both photocatalysts, absorption is almost uniform in UVB part, while reduces
approaching longer wavelengths in UVA part of spectrum (data from Figure 7. compared to
reported data [60-62]). The average values of absorbances in the UV-vis absorbance spectra
of TiO NT’s@Ag and TiO NT’s@Ag-HA compared to the average values for TiO NT’s
2
2
2
were used to determine ꢀ”app,310nm and ꢀ”app,365nm. All calculated values are reported in Table
1
.
4
. RESULTS AND DISCUSSION
16

ACCEPTED MANUSCRIPT
4
.1 Structural and optical characterization of TiO NT’s samples
2
Raman spectroscopy was applied as a method of choice to study the influence of the
decoration with Ag and hydrogenation on the structure of the obtained TiO nanotubes array.
2
Raman spectra of TiO NT’s, TiO NT’s@Ag and TiO NT’s@Ag-HA were shown in Figure
2
2
2
1
. The spectra of all the samples contained the observed bands assigned to five bands
-1
characteristics for anatase crystalline phase (144, 199, 396, 514 and 636 cm ) [63]. In the
-1
case of TiO NT’s@Ag-HA sample, the additional band at 246 cm (black arrow, Figure 1)
2
was observed due to annealing in a reductive atmosphere. The appearance of this band can be
explained by the existence of structural defects at the surface of nanotubes as results of
annealing in the reductive atmosphere [40, 64].
The scanning electron microscopy was applied to study the morphology of the samples
and influence of the modification and the decoration with Ag,. The SEM images show that
TiO NT’s obtained by simple anodization process have the inner diameter in the range of 55-
2
6
0 nm and an outer diameter in the range of 70-75 nm (Figure 2). It can also be observed that
Ag nanoparticles were uniformly deposited on the TiO nanotubes and the size distribution
2
are broad, which is preferable to improve efficiency due to the LSPR effect. Larger particles
(
size up to 100 nm) were seen at the surface of TiO NT’s, while smaller are distributed both
2
at the surface and inside the tubes as was additionally confirmed by TEM techniques (as will
be shown later). It could be noticed that the surface of TiO NT’s@Ag-HA appeared cleaner
2
than the surface of TiO NT’s@Ag due to annealing in the reducing gas atmosphere (Ar/H )
2
2
during which the adsorbed hydrocarbons and other impurities at the surface were removed.
The HAADF-STEM images of cross-section performed on the TiO NT’s@Ag sample,
2
shown in the Figure 3 a) and 3 b), indicate that TiO nanotubes have a length around 2100 nm
2
17

ACCEPTED MANUSCRIPT
and wall thickness between 14 and 20 nm. The specimens were anodized for 3 h and grew to
less than 3 µm in Fig. 3a. However, by using different electrolyte, nanotubes can grow to
more than 5 µm within 30 min of anodization [65, 66].
It was shown that Ag nanoparticles are uniformly distributed through the whole length of
nanotubes (Figure 3 a) and b)) in the average size of 5.7 nm for the Ag nanoparticles inside
the tubes (Figure 3 f)). The HRTEM analysis showed that nanotubes are in the anatase crystal
phase (Figure 3 c) and 3 d)) which is in agreement with Raman results (Figure 1). It could
also be observed in Figure 3 e) that Ag nanoparticles during photo deposition were embedded
into nanotube surface.
The TiO nanotube arrays were activated by the high-temperature annealing in the
2
reductive atmosphere, which results with partial amorphization and surface disorder of the
nanotubes surfaces (Figure 4) similar as in a case described in previous work [67]. The
amorphous layer on the TiO nanoparticles surface obtained by treatment of titanate
2
nanotubes (in form of powder) in hydrogen were also observed in our previous work [46] and
in previous studies with thermal treatment of titania nanostructures in the reduced atmosphere
[
38-49]. Here, it can be observed typical core-shell structure (crystalline core and thin
amorphous shell) of TiO nanotube from nanotubes array of sample TiO NT’s@Ag-HA. It is
2
2
3
+
expected that this layer would contain defects as oxygen vacancies and Ti states.
3
+
In order to confirm the presence of Ti in nanotubes amorphous shell after reduction, we
performed EELS spectroscopy measurements on nanotubes separated from the array by FIB.
The analysis of energy-loss near-edge fine structure (ELNES) of Ti-L2,3 edge and O k-edge
were carried out. As is known from the literature that Ti L-edge in case of anatase and rutile
core-loss EELS spectra consists of four major peaks (two doublets, t2g and e ) which
g
represent L and L edge, and splitted p orbitals, 2p and 2p1/2 states. It can be observed on
3
2
3/2
18

ACCEPTED MANUSCRIPT
the Figure 5 in ELNES of the Ti-L2,3 edge and the O-k edge of the TiO NT’s-HA and TiO2
2
NT’s@Ag-HA.
The spectra of both samples were recorder over the whole thickness of nanotube wall, and at
the surface of the nanotube. Comparing the spectra of the nanotube in core region and
surface, the EELS spectra of a Ti-L2,3 edge obtained at nanotube surface of both samples is
slightly shifted to lower energy loss values (Figure 5a and 5c) which imply the existence of
3
+
Ti electronic states on the nanotubes surface. Also, the EELS spectra obtained at nanotubes
surface looks similar to the spectra of amorphous TiO what is in agreement with previous
2
observations of order-disorder core-shell structure in the literature after high-temperature
treatment in the reductive atmosphere [67, 68]. On the Figure 5b and 5d are shown the EELS
spectra of O k-edge. It could be observed that there is almost no difference between spectra at
the core of nanotubes wall and surface.
The EELS spectra of pristine TiO NT’s and nanotubes treated in a reductive atmosphere,
2
TiO NT’s-HA and TiO NT’s@Ag-HA, were compared in Figure 6. The spectra of TiO
2
2
2
NT’s sample can be assigned to EELS spectra of the anatase phase. For the reduced samples
similar behaviour was observed (Figure 5). The splitting of Ti L and L -edge was less
2
3
pronounced in reduced samples. Additionally, spectra are slightly shifted to lower energy loss
3
+
due to creation of Ti states.
As reported recently [69], Pd nanoparticles can catalyse the reduction of TiO rutile in H2
2
at room temperature through the generation of highly active hydrogen species which interact
3
+
with TiO lattice and creates Ti , oxygen vacancies and even influencing the surface
2
disorder. Without the Pd, several days are necessary for the hydrogenation of the TiO , under
2
high pressure (20.0 bar) or temperature up to 700 °C. Surprisingly, the treatment of TiO2
matrix decorated with Ag nanoparticles did not improve the reduction. However, we
observed a successful reduction of TiO nanotube array with and without decoration with Ag,
2
19

ACCEPTED MANUSCRIPT
already after the samples were annealed for 3h at 500 � C using even less reductive
atmosphere (Ar/H ). Our observation of successful TiO hydrogenation can be explained by
2
2
nanotubular form of samples (which guarantee large surface of the material), but also by
anatase form of TiO nanotubes.
2
With the aim to study the influence of the decoration with Ag and hydrogenation on
broadening of the spectral region of the light absorption and narrowing of the TiO nanotubes
2
band gap, UV-vis measurements were performed. The results of UV-vis spectroscopy of the
TiO NT’s arrays show that in case of TiO NT’s-HA and TiO NT’s@Ag-HA, the
2
2
2
absorbance is shifted to the visible and NiR region of spectra (Figure 7a and 7b). The
bandgap of all studied samples was shown by the dashed line in the Figure 7c. As could be
observed the samples TiO NT’s and TiO NT’s-HA have similar main bandgap around 3.1
2
2
eV, but sample TiO NT’s-HA have additional transition around 1.6 eV. In the case of
2
samples TiO NT’s@Ag and TiO NT’s@Ag-HA there are following band gap 2.7 and 2.3
2
2
eV, respectively. The deposition of small Ag nanoparticles at the surface of nanotubes
significantly increase the absorption in the visible region of spectra. As it was already
mentioned in the introduction, in the case of plasmonic photocatalyst there is a strong
increase of absorption in the visible part of the spectra and partially in the UV region. The
same behaviour of absorption enhancement we observed in our measurements (Figure 7).
Annealing of TiO nanotubes decorated with Ag nanoparticles in the reductive atmosphere
2
Ar/H additionally influenced the structure and optical properties of materials. The annealing
2
in the reductive atmosphere created additional states in the gap between the conductive and
valence band of TiO , as is schematically shown in Figure 8, thus improving the absorption in
2
3
+
the visible region. During reduction process the structural defects are created, such as Ti , as
was proven by EELS measurements. Those defects influenced the electronic structure of
TiO nanotubes, which resulted in a macroscopic effect as narrowing of the bandgap. The
2
20

ACCEPTED MANUSCRIPT
created defects could be the traps for the recombination of created electron-holes pairs, but
due to the silver nanoparticles, this effect is suppressed.
In Figure 8, the schematic view of the influence of the silver on the optical and electrical
properties of TiO nanotubes are shown. On the Ag - TiO interface there is a creation of
2
2
Schottky barrier, which suppresses the recombination of light-induced charge carriers (e-h)
due to the internal local electric field at the Ag/TiO NTs interfaces. On the other hand, the
2
Ag nanoparticles induced LSPR effect which contributed to the final concentration of free
electron in the conduction band of TiO which can participate in the photocatalytic reaction.
2
The LSPR is size dependent effect so different sizes of the Ag nanoparticles obtained in our
work absorbed different Vis and NIR wavelengths. Therefore, if the LSPR electrons have
enough energy to overcome the Schottky barrier at the interface, the electrons would directly
transport into the conduction band of TiO . W. Hou and S. B. Cronin [70] in the review paper
2
state that for the better combined effect of TiO and noble metal nanoparticles (Ag, Au) and
2
higher photocatalytic efficiency there is a requirement for doping of TiO to create lower
2
energy states inside the bandgap of TiO . They said that crucial factor to achieve a maximum
2
of plasmon enhancement is that spectral enhancement of the metal nanoparticles overlap with
the spectral absorption of the semiconductor photocatalyst and with the spectrum of the
illumination source. They also said that most of the electron which is created due to LSPR
effect don't have enough high energy to overcome the Schottky barrier due to low Fermi level
in the noble metal nanoparticles. Noble metal nanoparticles (Ag) due to the LSPR effect
creates intense local electric fields and previous simulation in the literature shown that
electric field intensity of local plasmonic ‘’hot spots’’ can reach as 1000 times that of the
incident electromagnetic field [71, 72]. Which means that the ‘’hot spots’’ region can have
electron-hole pairs generation 1000 higher than in the case of an incident electromagnetic
field. Thus, enhancement in the amount of photo-induced charge generation is happening
21

ACCEPTED MANUSCRIPT
locally in the TiO . This enhancement will be possible only if there are additional lower
2
energy states are created inside the bandgap of TiO [70-73]. In our case additional energy
2
3
+
states are formed due to the defects such as: Ti and oxygen vacancies in the TiO which
2
enables the light absorption below the bandgap of TiO Due to previously mentioned reasons,
2
which were proved theoretically and experimentally, one can conclude that combined effect
of noble metal nanoparticles at the surface of TiO NT’s and defects state inside of bandgap
2
TiO NT created by subsequently annealing in the reduced atmosphere leads to highest
2
phocatalytic performances as will be shown and highest absorption as it was already shown in
the Figure 7.
4
.2 Photocatalytic properties of TiO NT’s samples
2
Results for the photocatalytic oxidation of salicylic acid using full-spectrum solar
simulation irradiation source and only visible light are given in Fig. 9. The photocatalytic
activity is observed in the following range: TiO NT’s@Ag-HA > TiO NT’s@Ag > TiO
2
2
2
NT’ in both cases. The significantly higher photocatalytic activity of TiNT@Ag-HA under
visible irradiation compared to the other studied catalysts can be ascribed to its significant
absorption in the visible regions of solar irradiation due to the effects explained previously.
Some authors show that the nonmetal doping of TiO with elements such as nitrogen and fluorine
2
are responsible for the nanotubes' photoactivity under visible light [74, 75], but EELS measurements
(
not shown) of our samples indicated that neither nitrogen nor fluorine were inbuilt during the
nonotubes' fabrication process. Given experimental data points are the mean results of three
series of experiments. The photocatalytic activities of studied photocatalysts did not vary
during repeated experiments suggesting the reusability and photocatalytic stability of as-
prepared samples.
22

ACCEPTED MANUSCRIPT
Kinetics of photocatalytic degradation of salicylic acid under simulated solar irradiation
and visible light, are shown in Fig. 9a and 9b, respectively. Under simulated solar irradiation,
photochemical degradation of salicylic acid was observed (Fig. 9a), which was expected
since SA contains chromophores that absorb at wavelengths 290-315 nm making it
susceptible to direct photolysis [76]. The photocatalytic degradation of SA is much faster
when TiO NT’s@Ag and TiO NT’s@Ag-HA were used as photocatalysts, disregarding the
2
2
source of irradiation (Fig. 9). TiO NT’s@Ag-HA film seems to be an efficient photocatalyst
2
even only under visible irradiation.
Regarding SA degradation intermediates and overall mineralization estimated based on
TOC measurements, results are given in Figure 10a. Results were compared using the C atom
balance. The theoretical value of TOC ascribed to the aromatic compounds i detected by
HPLC were calculated from Eq. (9), using a number of C atoms in aromatic compounds
N (C), molar masses (M ) and mass concentration (c ) for each compound.
i
i
i,m
1
2.01N (C)
i
TOC
= ∑
c
i,m
aromatics
M
i
i
(9)
The conversion of aromatic intermediates was higher than the achieved mineralization
extent. Expected aliphatic intermediates, i.e. short-chain carboxylic acids cannot be readily
degraded towards CO which accounts for higher residual TOC values. The mineralization
2
efficiency was highest when TiO NT’s@Ag-HA film was applied, for both irradiation
2
sources. Mineralization under visible light is negligible, however the percentage of residual
aromatics is lower after 120 min of photocatalytic reaction, pointing at the slow ongoing
conversion of aromatics towards aliphatic compounds.
In the preliminary analysis, kinetics of SA photocatalytic degradation was evaluated
according to the pseudo-first order reaction rate model in batch reaction system (Eq. (10)):
dc
SA
dt
=
−k
c
SA, obs SA
(10)
23

ACCEPTED MANUSCRIPT
where c is a concentration of salicylic acid, while kSA, obs is the pseudo-first order rate
SA
-1
constant in min , and it was evaluated through the linear regression of ln(c /cSA,0) vs. time
SA
(Eq. (11)).
c
SA
ln
=−k t
SA, obs
c
SA,0
(11)
Corresponding values of kSA, obs, degradation half-time, t1/2 and determination coefficient
2
R are given in Table 2. Based on experimental results, adsorption of SA on films was
insignificant. The pseudo-first order rate constant for the direct photolysis under simulated
-3
-1
solar irradiation of aqueous salicylic acid was found to be ≈ 2.0 × 10 min . The obtained
-4
-1
-1
value corresponds well with the reported value of < 3 × 10 min (0.02 hr ) [68] having in
mind that the peak intensity of applied irradiation in UVB region (310 nm) is almost 10 times
higher than the average intensity of UVB irradiation accounted for natural sunlight .
Repeated experiments were done in a row to test the stability of prepared photocatalytic
films: TiO NT’s, TiO NT’s@Ag, TiO NT’s@Ag-HA under simulated solar irradiation.
2
2
2
The resulting kSA, obs of sets of 10 experiments (cycles) are shown in Fig. 10b . The uniformity
-1
of the pseudo-first order rate constant for SA degradation (kSA, obs, min ) is observed when
using TiO NT’s, TiO NT’s@Ag, TiO NT’s@Ag-HA as photocatalysts, leading to a
2
2
2
conclusion that prepared films are equally active and thus stable for at least 20 hours.
The preliminary kinetics study based on pseudo-first order reaction rate constant was
insufficient to describe underlying phenomena. The detailed kinetic model was developed to
include the contribution from direct photolysis as well as the contribution from UVB, UVA
and visible light to photocatalytic degradation of SA (see section 3.; Eq. (1)).
The main hypothesis in this study was the achievement of advantageous photocatalytic
properties of TiO NT’s@Ag and TiO NT’s@Ag-HA compared to anatase film. Tentatively,
2
2
the photocatalytic properties of these films would exceed the properties of the efficient
commercial TiO , i.e. TiO P25 in the case when the same would be prepared in a form of
2
2
24

ACCEPTED MANUSCRIPT
thin film, hypothetically, as described in Section 3. Calculated values of model parameters
are reported in Table 1. Comparison of kinetic models and experimental results are shown in
Fig. 11. A simple model (Eq. (10)) is in accordance with the developed complex model (Eqs.
(1), (2) and (8)).
The novelty within this work is the overall factor of synergy (f ). This factor summarized
Syn
the influences of the combined effect of local surface plasmon resonance of Ag and creation
3
+
of amorphous layer with Ti and oxygen vacancies states at nanotubes surfaces (responsible
for the decrease of E and absorption in the visible part of the spectrum). The f includes all
g
Syn
underlying phenomena which resulted in higher rates of radical species generation on active
-
sites, such as OH•, O • etc. The f were given for all studied photocatalytic films (Table 1).
2
Syn
The value of 1.13 was obtained for TiO NT’s@Ag, while it was 2.56 for TiO NT’s@Ag-
2
2
HA. Such difference suggests that the film obtained by annealing in reductive atmosphere
indeed poses additional states in the gap between the conductive and valence band which is
vis
3+
also confirmed by determined Pph , and the structural defects are created, such as Ti and
oxygen vacancies, which influence on the electronic structure of TiO nanotubes and account
2
for overall synergy. Using the presented kinetic modeling approach (Eq. (8)) meaningful
values of apparent relative absorption coefficients are presented (ꢀ”). Enhanced light
absorption throughout the solar spectra can be concluded form the following: (i) the
ꢀ
”app,310nm and ꢀ”app,365nm calculated for TiO NT’s@Ag and TiO NT’s@Ag-HA are higher
2
2
vis
ph
than those assumed for anatase film and TiO P25, (ii) the P
values significantly increased
2
in the following order: TiO NT’s < TiO NT’s@Ag < TiO NT’s@Ag-HA. The results from
2
2
2
kinetic modeling supported by presented experimental results confirmed the conclusions
made entirely based on the structural and optical characterization of the film samples.
25

ACCEPTED MANUSCRIPT
5
. CONCLUSIONS
The influence of high-temperature annealing of TiO nanotube arrays decorated with silver
2
nanoparticles in the reductive Ar/H atmosphere on the photocatalytic activity was studied.
2
Three different samples of TiO nanotubes were prepared and studied: pure TiO nanotube
2
2
arrays, TiO nanotubes array where interior and exterior of all the nanotubes are decorated by
2
silver nanoparticles (TiO NT’s@Ag), and TiO NT’s@Ag treated at 500 °C in a hydrogen
2
2
atmosphere (TiO NT’s@Ag-HA). The synthesis of nanotubes by anodization of Ti foils is
2
simple, fast and inexpensive as well as further deposition of Ag and treatment in hydrogen.
The absorbance of solar irradiation was extended into the visible region for the samples
TiO NT’s@Ag and TiO NT’s@Ag-HA. Surface plasmon resonance of deposited Ag
2
2
nanoparticles was responsible for the increase of absorption in the case of TiO NT’s@Ag. In
2
sample TiO NT’s@Ag-HA, in addition to surface plasmon resonance, the defects created by
2
hydrogen treatment generate additional transition state inside the bandgap of TiO and further
2
increase absorbance.
The efficiency of photocatalysis was tested on salicylic acid, recognized as a typical
environmental pollutant. The obtained results show much better photocatalytic performance
of TiO NT’s@Ag than commercial TiO P25, while the best performance was observed for
2
2
TiO NT’s@Ag-HA.
2
The detailed kinetic study showed the contribution from photon absorption effects in
UVB, UVA and visible part of the simulated solar spectrum. Results showed the enhanced
vis
absorption of irradiation through the spectrum, in comparison with P25. The Pph values
indicate the well-enhanced absorption rates in the following range TiO NT’s < TiO2
2
NT’s@Ag < TiO NT’s@Ag-HA. The results from kinetic modeling supported the
2
understanding of the influence of a visible part of the solar spectrum to overall photocatalytic
26

ACCEPTED MANUSCRIPT
performances. The photocatalytic effect is enlarged by the synergy of absorbance in UV part
(
in bulk of TiO NTs) and visible part of solar irradiation (explained by combined effects of
2
3
+
surface plasmon resonance of Ag and creation of amorphous layer with Ti on the surface of
TiO NTs). The determined factors of synergy (f ) for TiO NT’s@Ag and TiO NT’s@Ag-
2
Syn
2
2
HA were 1.36 and 2.71, respectively.
The additional substantial advantage of studied photo-catalysts compared to the TiO2
nanoparticles based materials (as P25) is easy handling. They are formed and completely
attached on titanium plates that can be easily removed from the solution after the
photocatalytic reaction. Since the rate of photocatalytic degradation of model pollutant is
quite high when solar simulation was applied (less than 2 h is needed for its half-degradation
if TiO NT’s@Ag-HA film was used), this gives a certain merit to a possible use of this type
2
of photocatalysts on a larger scale. Obtained results are promising in the way of designing
efficient reactors for the future application in the field of waste-water purification using TiO2
membranes or by constructing reactor walls made of TiO nanotubes array decorated by Ag
2
nanoparticles and reduced in the Ar/H atmosphere.
2
ACKNOWLEDGMENT
Funding: This work has been supported in part by Croatian Science Foundation under the
project IP-2014-09-9419, a Croatian-Germany bilateral project financed by DAAD agency
and Ministry of science and education of Croatia and Centre of Excellence for Advanced
Materials and Sensing Devices, Ruđer Bošković Institute, Zagreb, Croatia.
27

ACCEPTED MANUSCRIPT
REFERENCES:
1] B.O’Regan, and M. Grätzel, A Low-cost, High-Efficiency Solar Cell Based on Dye-Sensitized
Colloidal TiO Films, Nature 353 (1991) 737-740.
[
2
[
[
2] M. A. Fox, and M. T. Dulay, Heterogeneous Photocatalysis, Chem. Rev. 93 (1993) 341-357.
3] A. Fujishima, K. Honda, Electrochemical Photolysis of Water at a Semiconductor Electrode,
Nature 238 (1972) 37-38.
4] W. Guo, X. Xue, S. Wang, C. Lin, Z. L. Wang, An Integrated Power Pack of Dye-Sensitized Solar
Cell and Li Battery Based on Double-Sided TiO Nanotube Arrays, Nano Lett. 12 (2012) 2520-2523.
[
2
[
[
5] R. Van Noort, Titanium: The Implant Material of Today, J. Mat. Sci. 22 (1987) 3801-3811.
6] X. X. Zhang, J.B. Zhang, Y.C. Jia, P.Xiao, J. Tang, A Pt-Doped TiO Nanotube Arrays Sensor for
2
Detecting SF Decomposition Products, Sensors 12 (2012) 3302–3313.
6
[
7] D. Gong, C. A. Grimes, O. K. Varghese, W. Hu, R. S. Singh, Z. Chen, and E. C. Dickey, Titanium
Oxide Nanotube Arrays Prepared by Anodic Oxidation, J. Mater. Res. 16 (2001) 3331-3334.
8] K. Lee, A. Mazare, P. Schmuki, One-Dimensional Titanium Dioxide Nanomaterials: Nanotubes,
Chem. Rev. 114 (19) (2014) 9385–9454.
9] D. Yu, X. Zhu, X. Zhen, X. Zhong, Q. Gui, S. Ye, S. Zhang, X. Chen, D. Li, Facile method to
enhance the adhesion of TiO2 nanotube arrays to Ti substrate, ACS Applied Materials & Interfaces 6
[
[
(
[
2014) 8001-8005.
10] F. Fabregat-Santiago, E. M. Barea, J. Bisquert, G. K. Mor, K. Shankar, C. A. Grimes, High
Carrier Density and Capacitance in TiO Nanotube Arrays Induced by Electrochemical Doping, J.
2
Am. Chem. Soc. 130 (2008) 11312–11316.
[
11] J.-G. Li, T. Ishigaki, and X. Sun, Anatase, Brookite, and Rutile Nanocrystals via Redox
Reactions under Mild Hydrothermal Conditions:ꢀ Phase-Selective Synthesis and Physicochemical
Properties, J. Phys. Chem. C 111 (2007) 4969-4976.
[
12] I. Paramasivam, H. Jha, N. Liu, P. Schmuki, A review of photocatalysis using self-organized
TiO nanotubes and other ordered oxide nanostructures, Small 8 (2012) 3073-3103.
2
28