TiO2-Bi2O3/(BiO)2CO3-reduced graphene oxide composite as an effective visible light photocatalyst for degradation of aqueous bisphenol A solutions

Article abstract of DOI:10.1016/j.cattod.2018.02.039

TiO₂ nanorods (T) were combined with a narrow band gap semiconductor β-Bi₂O₃ (B) to form a heterojunction, which makes it possible for TiO₂ to become active as a photocatalyst also under visible light illumination. To further increase the photocatalytic activity of TiO₂ + Bi₂O₃/(BiO)₂CO₃ (TB) composite, we used a hydrothermal procedure to link it with reduced graphene oxide (rGO). Structural, surface and electronic properties of the obtained catalysts were analyzed and correlated to their performance in photocatalytic oxidation of aqueous bisphenol A (BPA) solution conducted in a batch reactor under visible light illumination. XRD, FTIR, UV–vis DR spectroscopy and photocurrent measurements of visible light illuminated TB composite catalyst clearly showed that (i) β-Bi₂O₃ acts as a photosensitizer for TiO₂ and (BiO)₂CO₃ present in the TB composite, (ii) holes (h⁺) are photo-generated in valence band (VB) of β-Bi₂O₃ and due to the β-Bi₂O₃/TiO₂ heterojunction transferred into VB of TiO₂, (iii) p-n junction between β-Bi₂O₃ and TiO₂ allows the photo-generated electrons (e^?) in the conduction band (CB) of β-Bi₂O₃ to transfer to TiO₂, and (iv) p-n junction between β-Bi₂O₃ and (BiO)₂CO₃ allows the photo-generated electrons in the conduction band of β-Bi₂O₃ to transfer to (BiO)₂CO₃. This means that more charge carriers are available to participate in the catalytic visible-light triggered oxidation process for the degradation of organic pollutants dissolved in water. The highest photocurrent density was measured for multi-phase TBR (TB + rGO) composite, which indicates that visible-light generated charge carriers in TB composite are injected into the reduced graphene oxide. The latter acts as a web for charge carrier percolation and suppresses the recombination of electron-hole pairs, thus resulting in improved catalytic activity of TBR. The results of UV–vis DR spectroscopy and photocurrent density measurements were entirely in line with the results of photocatalytic oxidation of water dissolved bisphenol A (BPA) in batch reactor under visible light illumination.

Full text of DOI:10.1016/j.cattod.2018.02.039

Accepted Manuscript
Title: TiO2-Bi2O3/(BiO)2CO3-reduced graphene oxide
composite as an effective visible light photocatalyst for
degradation of aqueous bisphenol A solutions
ˇ
´
Authors: Gregor Zerjav, Petar Djinovic, Albin Pintar
PII:
S0920-5861(18)30111-1
DOI:
Reference:
https://doi.org/10.1016/j.cattod.2018.02.039
CATTOD 11264
To appear in:
Catalysis Today
Received date:
Revised date:
Accepted date:
3-11-2017
14-2-2018
22-2-2018
ˇ
´
Please cite this article as: Gregor Zerjav, Petar Djinovic, Albin Pintar, TiO2-
Bi2O3/(BiO)2CO3-reduced graphene oxide composite as an effective visible
light photocatalyst for degradation of aqueous bisphenol A solutions, Catalysis
Today https://doi.org/10.1016/j.cattod.2018.02.039
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.

TiO2-Bi2O3/(BiO)2CO3-reduced graphene oxide composite as an effective visible light
photocatalyst for degradation of aqueous bisphenol A solutions
Gregor Žerjav*, Petar Djinović, Albin Pintar
Department for Environmental Sciences and Engineering, National Institute of Chemistry,
Hajdrihova 19, SI-1001 Ljubljana, Slovenia.
*Corresponding author. Tel.: +386 1 47 60 249; fax: +386 1 47 60 460. E-mail address:
gregor.zerjav@ki.si (G. Žerjav).
GRAPHICAL ABSTRACT
A novel multi-junction TBR photocatalyst (containing TiO2, β-Bi2O3/(BiO)2CO3 and reduced
graphene oxide) was prepared for oxidation of water dissolved bisphenol A (BPA) in advanced
oxidation process under visible light illumination. The graphic shows band positions of each
component before their close chemical contact as well as possible migration pathways of
electrons and holes.
1

Highlights






- Synthesis of multicomponent visible light active catalyst TiO2-Bi2O3/(BiO)2CO3-rGO.
- Bi2O3 act as photosensitizer for TiO2 and (BiO)2CO3.
- Heterojunction TiO2/Bi2O3 enables the transfer of photo-generated holes.
- P-n junction TiO2/Bi2O3 enables the transfer of photo-generated electrons.
- P-n junction Bi2O3/(BiO)2CO3 enables the transfer of photo-generated electrons.
- Reduced graphene oxide (rGO) acts as a web for the percolation of charge carriers.
Abstract - TiO2 nanorods (T) were combined with a narrow band gap semiconductor β-Bi2O3
(B) to form a heterojunction, which makes it possible for TiO2 to become active as a
photocatalyst also under visible light illumination. To further increase the photocatalytic
activity of TiO2+Bi2O3/(BiO)2CO3 (TB) composite, we used a hydrothermal procedure to link
it with reduced graphene oxide (rGO). Structural, surface and electronic properties of the
obtained catalysts were analyzed and correlated to their performance in photocatalytic oxidation
of aqueous bisphenol A (BPA) solution conducted in a batch reactor under visible light
illumination. XRD, FTIR, UV-Vis DR spectroscopy and photocurrent measurements of visible
light illuminated TB composite catalyst clearly showed that (i) β-Bi2O3 acts as a photosensitizer
for TiO2 and (BiO)2CO3 present in the TB composite, (ii) holes (h+) are photo-generated in
valence band (VB) of β-Bi2O3 and due to the β-Bi2O3/TiO2 heterojunction transferred into VB
of TiO2, (iii) p-n junction between β-Bi2O3 and TiO2 allows the photo-generated electrons (e-)
in the conduction band (CB) of β-Bi2O3 to transfer to TiO2, and (iv) p-n junction between β-
Bi2O3 and (BiO)2CO3 allows the photo-generated electrons in the conduction band of β-Bi2O3
to transfer to (BiO)2CO3. This means that more charge carriers are available to participate in
the catalytic visible-light triggered oxidation process for the degradation of organic pollutants
dissolved in water. The highest photocurrent density was measured for multi-phase TBR
(TB+rGO) composite, which indicates that visible-light generated charge carriers in TB
composite are injected into the reduced graphene oxide. The latter acts as a web for charge
carrier percolation and suppresses the recombination of electron-hole pairs, thus resulting in
2

improved catalytic activity of TBR. The results of UV-Vis DR spectroscopy and photocurrent
density measurements were entirely in line with the results of photocatalytic oxidation of water
dissolved bisphenol A (BPA) in batch reactor under visible light illumination.
Keywords: visible light active catalyst; photocatalysis; TiO2-Bi2O3/(BiO)2CO3-reduced
graphene oxide composites; charge carriers migration; bisphenol A; reduced graphene oxide;
TiO2 nanorods; Bi2O3
1. Introduction
Due to its high photocatalytic activity, chemical stability, low cost and nontoxicity, titanium
dioxide (TiO2) holds several promising applications as a photocatalyst [1,2]. TiO2 is one of the
most suitable catalysts for heterogeneous photocatalytic oxidation of many organic pollutants
present in aqueous medium due to its high-lying conduction band (CB) [3]. If a photocatalyst
is illuminated by light with energy equal to, or greater than its band gap energy, an electron (e-
) in the valence band (VB) moves to the conduction band and leaves a hole (h+) in the VB.
Photo-generated holes and electrons can either (i) reduce and oxidize substrates adsorbed on
the catalyst surface, or (ii) recombine with each other without any chemical reaction.
Recombination of electrons and holes hinders the photocatalytic activity of catalysts. The
overall efficiency of a photocatalyst is determined by its ability for (i) light absorption, (ii)
electron-hole pair separation and migration, and (iii) efficiency of electron-hole pair utilization.
The main drawbacks of TiO2 as a photocatalyst are (i) band gap energy of 3.2 eV, which means
that it can be excited only by ultraviolet (UVA) light (λ<387 nm), and (ii) fast electron-hole
recombination.
In our previous study [4], we decorated anatase TiO2 nanorods with reduced graphene oxide
(rGO) to obtain TiO2+rGO composite. Its structural, surface and electronic properties [5]
revealed improved charge separation compared to pure TiO2 due to the perfect matching of
TiO2 and rGO valence band maxima (VBM). The excited electrons of TiO2 nanorods are
3

injected into the CB of rGO due to band alignment, which minimizes the radiative electron-
hole recombination. The rGO acts as a web for electron percolation, which significantly
improves the electron migration [6,7]. It was also shown that it is critical to optimize the amount
of rGO in the TiO2+rGO composites to achieve the highest photocatalytic activity. Excess rGO
can promote charge carrier recombination, thus decreasing the number of electrons and holes
participating in a catalytic reaction [8,9]. Combining TiO2 and rGO in a proper ratio presents a
successful solution to hinder electron-hole recombination in TiO2 and increase its
photocatalytic activity [10-]. The downside of wide band gap energy of TiO2 still remains,
meaning that only UV light can trigger the excitation of electrons and formation of holes in
such solids.
There are several options available to extend the light response of TiO2 into the visible light
region [13]. Coupling TiO2 with a narrow band gap semiconductor (WO3, In2O3, SnO2, ...) is
an effective way to improve the photostimulation and electron-hole separation. Bismuth based
compounds were recently very effectively used to form a heterostructure and improve the
optical response of TiO2 (BiOI/TiO2, BiOCl/TiO2, BiVO4/TiO2, ...) [14-]. Bi2O3 has four
different polymorphs, namely: monoclinic (α), tetragonal (β), body-centered cubic (γ) and face-
centered cubic (δ), all exhibiting notably different light-matter interactions. The band gap
energy of β-Bi2O3 (2.4 eV) is notably lower than that of α-Bi2O3 (2.8 eV), resulting in light
absorption in a wider visible-light region. This manifests itself in β-Bi2O3 polymorph showing
higher photocatalytic activity compared to α-Bi2O3 in the decomposition of methyl orange and
p-chlorophenol under visible light illumination [17-]. Therefore, β-Bi2O3 appears as a
promising candidate to form a heterojunction with TiO2 and to use its photosensitizing effect to
boost visible-light catalytic performance of TiO2. The photocatalytic activity of pure β-Bi2O3
is poor because of the following unfavorable properties: (i) conduction band potential of Bi2O3
-
is too low to oxidize the surface O2 to O2 through a fast single-electron reaction, resulting in
4

fast electron-hole recombination [20], (ii) synthesis procedures of pure β-Bi2O3 generally
produce very large particles and consequently low specific surface area materials (below 1
m2/g) [21], and (iii) narrow band gap favors electron-hole recombination [22]. Several authors
report that when β-Bi2O3 (p-type semiconductor) and (BiO)2CO3 (n-type semiconductor) are
combined, a p-n heterojunction is formed which improves the photocatalytic activity of the
composite. In this composite, the role of β-Bi2O3 is to act as a visible-light photosensitizer for
(BiO)2CO3, which has a large band gap of 3.1 eV and is not photocatalytically active under
visible light irradiation [23-]. The fast electron-hole recombination in β-Bi2O3 is hindered due
to the transfer of photo-generated charge carriers to (BiO)2CO3, resulting in improved visible
light catalytic activity of β-Bi2O3.
The aim of this work was to synthesize innovative multicomponent TiO2-Bi2O3/(BiO)2CO3-
rGO composites, where the role of: (i) Bi2O3 is to act as a visible light photosensitizer, (ii)
appropriately positioned VB and CB of TiO2 and (BiO)2CO3 act as selective sinks for photo-
generated holes and electrons, respectively, and (iii) rGO acts as a web for the percolation of
charge carriers that prolongs their separation and lifetime. We investigated how different
combinations of TiO2, -Bi2O3, (BiO)2CO3 and rGO influence the structural and electronic
properties of the synthesized composites, and how this manifests itself in the photocatalytic
activity towards liquid-phase bisphenol A degradation. Bisphenol A (BPA) is an endocrine
disrupting compound which finds widespread use in production of mostly packaging materials
used every day [29]. These materials end as a waste in landfills, which usually leads to slow
leaching of BPA into ground water. Therefore, due to environmental implications BPA was
chosen as a model compound to test the photocatalytic activity of developed photocatalysts.
2. Experimental
2.1. Sample preparation
2.1.1. Synthesis of TiO2 nanorods and graphene oxide
5

TiO2 nanoparticles (DT-51, provided by Crystal Company) were hydrothermally treated to
obtain TiO2 nanorods, which were further calcined at 500oC in air for 2 h to obtain pure anatase
TiO2 nanorods (sample denoted as T). The details about the synthesis procedure are available
in our previous work [4], where also the details about the synthesis procedure of graphene oxide
are described (sample denoted as GO).
2.1.2. Synthesis of bismuth oxide (B) and TB composite
Solution combustion method was used to prepare pure bismuth oxide (sample denoted as B)
and TiO2-Bi2O3 composite (denoted as TB). To prepare Bi2O3, 2.9 g of Bi(NO3)3∙5H2O
(Honeywell Fluka) and 1.4 g of citric acid monohydrate (C6H8O7∙H2O, Merck) were dissolved
in 10 ml HNO3 (Merck, 0.04 M) and stirred for 1 h. After complete dissolution, 0.04 g Pluronic®
P-123 (Sigma-Aldrich) was added and stirred for 4 h. The obtained solution was transferred
into a ceramic cup and placed inside an oven, where it was heated (temperature ramp 120oC/h)
to 300oC for 24 h and left to cool down naturally to room temperature. In the case of TB
composite synthesis, 0.58 g of Bi(NO3)3∙5H2O (Honeywell Fluka) and 0.29 g of citric acid
monohydrate (C6H8O7∙H2O, Merck) were dissolved in 10 ml HNO3 (Merck, 0.04 M) and stirred
for 1 h. After that, 0.008 g of Pluronic® P-123 (Sigma-Aldrich) was added and stirred for
another hour. Finally, 1 g of TiO2 was added and stirred for 3 h. The suspension was transferred
into a ceramic cup and placed inside an oven, where it was heated to 300oC for 24 h (temperature
ramp 120oC/h) and allowed to cool down naturally to room temperature. The nominal
composition of the TB composite was 0.25 g of Bi2O3 and 1 g of TiO2.
2.1.3. Synthesis of TR, BR and TBR composites
Graphene oxide (0.05 g) was added to a mixture of absolute ethanol (Sigma Aldrich, 25 ml)
and deionized water (50 ml). The amount of GO and ultimately rGO in the composites (4 wt.
%) was chosen based on the experience gained during our previous study [4]. To exfoliate GO,
the mixture was ultrasonicated for 2 h. Afterwards, 0.5 g of T, B or TB powders were added to
6

the suspension which was ultrasonicated for another 2 h. The suspension was transferred into a
200 ml Teflon-lined autoclave and heated to 130oC for 4 h to hydrothermally reduce GO. The
resulting products were recovered by centrifugation, washed with deionized water and dried in
vacuum under cryogenic conditions. Samples were denoted as TR (TiO2+rGO composite), BR
(Bi2O3+rGO composite) and TBR (TB+rGO composite).
2.2. Characterization
Specific surface area, total pore volume and average pore size were determined using N2
adsorption/desorption isotherms at -196oC (Micromeritics, model TriStar II 3020). The samples
were before measurements degassed using a SmartPrep degasser (Micromeritics) in N2 stream
(Linde, purity 6.0) at elevated temperature (60 min at 90oC, followed by 240 min at 180oC).
Brunauer, Emmett and Teller (BET) theory was applied in order to calculate the specific surface
area of composites.
The morphology of synthesized materials was examined by field-emission scanning electron
microscope (Carl Zeiss, model FE-SEM SUPRA 35VP), equipped with energy-dispersive
detector (Oxford Instruments, model Inca 400).
X-ray powder diffraction (XRD) patterns of the catalysts were collected on PANanalytical
X`pert PRO MPD diffractometer with Cu Kα1 radiation (1.54056 Å) in reflection geometry
(scan range: 20-90o in increments of 0.034o). PDF standards from the International Centre for
Diffraction Data (ICDD) were used for the identification of crystalline phases.
UV-Vis diffuse reflectance (UV-Vis DR) spectra were obtained at room temperature using
Perkin-Elmer Lambda 35 UV-Vis spectrophotometer equipped with the RSA-PE-19M Praying
Mantis accessory. White reflectance standard Spectralon© was used to record the background
correction in the range of 200-700 nm.
FTIR measurements were recorded using the Perkin-Elmer FTIR spectrometer (model
Frontier). Sample pellets were prepared by mixing 1 mg of the sample with 99 mg of KBr. They
7

were ground in an agate mortar and pressed at 5 tonnes for 5 minutes. The spectra were collected
in transmission mode using the MCT detector. The spectra shown are the average of 16 scans
with a resolution of 4 cm-1 in the spectral range of 4000-450 cm-1.
The photoresponse characteristics of prepared materials were evaluated by Metrohm Autolab
PGSTAT30 potentiostat/galvanostat in a three-electrode electrochemical cell under intermittent
visible light illumination (halogen lamp 150 W, λmax=ꢀ520 nm, UV cut-off filter at 410 nm,
Fig. S1) with 0ꢀV bias potential (vs. SCE) in 0.1 M KOH aqueous solution. The filter was used
to ensure that the working electrodes with deposited samples were illuminated only by visible
light (Fig. S1 shows UV-vis DR spectra of the cut-off filter and the halogen lamp).
Electrochemical impedance (EIS) spectra of prepared catalysts were obtained in the frequency
range of 0.1-106 Hz; 0.1 M Na2SO4 was used as electrolyte. Screen-printed DropSens electrode
(DropSens DRP-150) was used as working electrode, platinum as a counter electrode and
calomel electrode (HANNA instruments HI5412) as a reference electrode. To prepare the
working electrode, 10 μl of catalyst-ethanol suspension (12.5 mg of catalyst dispersed in 2.5 ml
of absolute ethanol (Sigma Aldrich)) was dropped onto the surface of the screen-printed
electrode and dried overnight at room temperature.
2.3. Photocatalytic oxidation experiments
A solution of bisphenol A (c0=10 mg/l, BPA) in ultrapure water (18.2 MΩcm) was used for the
photocatalytic experiments. The experiments were performed in a 250 ml batch slurry reactor
(Lenz, model LF60, 250 ml) at atmospheric pressure, which was magnetically stirred (600 rpm)
and thermostated at 20oC (Julabo, model F25/ME). Solution was purged with purified air during
the whole experiment (45 l/h). The catalyst concentration used in all experiments was 125 mg/l.
Ultrasonication was employed to suspend the catalyst before adding to the BPA solution. The
suspension was kept in dark for 30 min (“dark” period, for establishing of the sorption process
equilibrium) before illumination with halogen lamp (150 W, λmax=ꢀ520 nm). An UV cut-off
8

filter at 410 nm (Rosco E-Colour #226: U.V. filter) was used in all experiments to ensure that
the catalysts tested were illuminated only by visible light (Fig. S1). The halogen lamp with
UV filter was positioned in a water-cooled quartz jacket, which was immersed vertically in the
center of the batch slurry reactor.
2.4. Analysis of end-product solutions
To determine temporal BPA conversions, 1.5 ml samples were withdrawn in 5-30 min intervals
during the photolytic/photocatalytic runs. The samples were filtered through a 0.2 μm
membrane filter and analyzed with an HPLC instrument (Thermo Scientific, model Spectra).
Details about the HPLC analytical protocol can be found elsewhere [30]. By measuring the total
organic carbon (TOC) content at the end of photocatalytic experiments using an instrument
from Teledyne Tekmar (model Torch), the level of mineralization, i.e. the total amount of
removed organic substances in aqueous-phase samples was determined.
3. Results and discussion
3.1. Catalyst characterization
3.1.1. SEM-EDX and N2 physisorption analyses
SEM analysis was utilized to visualize the morphology of prepared solids (Fig. 1). A proper
rod-like morphology was observed for the T sample (Fig. 1) with a length of 80±8 nm and a
diameter of 10±2 nm. The morphology of pure Bi2O3 (B) was shaped as nanoplates with
thickness of about 300 nm. Differences in the shape of B and T morphologies are expressed
also in their BET surface areas (Table 1), where values of 4.7 and 105 m2/g were calculated for
samples B and T, respectively. Corresponding N2 adsorption-desorption isotherms and BJH
pore size distributions are present in Fig. 2. Low BET surface area of sample B was observed
also by others [21]. After exfoliation and reduction of GO, a sheet-like morphology of sample
rGO was obtained. SEM images (Fig. 1) reveal that after combining pure T with rGO or/and B,
the nanorod-like structure of T is maintained, which implies that its morphology is not
9

influenced by the subsequent hydrothermal GO reduction. In Fig. 3, SEM-EDX elemental
mapping of TBR composite shows that bismuth is well distributed in the solid and decorated
with titanium. Also, no segregation or clustering of T or B components could be noticed. Carbon
was omitted from the results of SEM-EDX elemental mapping due to the fact that the samples
were fixed with carbon tape, so we couldn’t reliably distinguish between carbon originating
from the samples or carbon tape.
3.1.2. XRD analysis
Individual components (T, B and R) as well as their multi-phase composites were analyzed
using XRD technique (Fig. 4). All diffraction lines observed in the XRD pattern of pure B
correspond to tetragonal (β) Bi2O3 (JCPDS 00-27-0050). The peak at 9.8o in the pure GO
sample corresponds to the (002) interlayer spacing (d) of 0.90 nm, which matches very well
with the value reported for graphene oxide [31]. In all other materials containing rGO, these are
not visible, indicating successful reduction of GO to rGO. Analysis of pure T sample revealed
only diffraction peaks characteristic for anatase TiO2.
During analysis of composites, drastic changes related to -Bi2O3 structure were observed. T
appears to be not affected in any of the materials: it remains as anatase and its average scattering
domain size changes negligibly (14, 15, 16 and 15 nm for pure T, TB, TR and TBR samples,
respectively; see Table 1). In the TB composite, tetragonal -Bi2O3 is present only as a minor
phase (based on the intensity of its diffraction lines). The main crystalline bismuth containing
component is bismuth carbonate ((BiO)2CO3, JCPDS 00-041-1488). In BR composite,
tetragonal -Bi2O3 is the prevailing phase in coexistence with bismuth carbonate ((BiO)2CO3).
Based on these results, we can postulate that presence of anatase T strongly hinders formation
of -Bi2O3 during calcination step and the predominant bismuth-containing phase remains
(BiO)2CO3. It is known that -Bi2O3 cannot be produced by thermal decomposition of bismuth
nitrate (more stable - and -Bi2O3 polymorphs are produced in this case), but could be instead
10

readily formed from bismuth carbonate [27]. Utilization of calcination temperatures higher than
300oC results in formation of α-Bi2O3 polymorph [27], which has a broader bandgap and
consequently is unsuitable as a visible-light sensitizer of TiO2. Also, -Bi2O3 is not stable under
the employed hydrothermal conditions used for GO reduction, leading to its partial conversion
to (BiO)2CO3.
In the TBR composite, the only crystalline phase observed was anatase and only traces of -
Bi2O3 and (BiO)2CO3 were observed. Bismuth oxide is predominantly present in an amorphous
form. This conclusion is supported by the fact that SEM-EDX elemental analysis shows
bismuth content very close to the nominal one. Actual chemical compositions of TB and TBR
composites calculated based on the SEM-EDX results are: 0.28 g Bi2O3 and 1 g TiO2, and 0.21
g Bi2O3 and 1 g TiO2, respectively.
3.1.3. FTIR analysis
Acquired FTIR spectra of prepared single-component catalysts are presented in Fig. 5a. The
spectrum of sample T shows strong absorption in the range of 1000-450 cm-1, which
corresponds to lattice Ti-O-Ti vibrations. The broad band in the range of 3700-2500 cm-1 and
a peak at 1640 cm-1 can be ascribed to stretching and bending vibrations of surface adsorbed
hydroxyl groups and water, respectively [32].
The bands in the spectrum of GO sample appearing at 1050, 1220, 1414 and 1720 cm-1 are
stretching vibrations of oxygen containing groups (alkoxy C-O, epoxy C-O, C-OH and carbonyl
C=O). The broad band appearing between 3700 and 2500 cm-1 arises from stretching vibrations
of surface hydroxyl (-OH) groups, while the band at 1626 cm-1 is attributed to C=C vibration
of graphene oxide, thus confirming its sp2 hybridized structure [33,34]. The latter is overlapping
with water vibrations expected at 1630-1640 cm-1.
In the B sample, bands below 700 cm-1 belong to the lattice Bi-O bond vibrations [35]. Strong
absorption bands of CO32- groups are observed at 1388 and 1440 cm-1. These can be assigned
11

to its anti-symmetric vibration ν3 mode. At 845 cm-1, an out of plane ν2 bending mode
characteristic of (BiO)2CO3 is observed. The broad absorption band between 3700 and 2700
cm-1 and at 1635 cm-1 can be attributed to the O-H stretching mode and bending modes of
adsorbed water molecules, respectively [36].
Fig. 5b shows FTIR spectra of composites. In the TB composite, an intense Ti-O-Ti band
(between 450 and 1000 cm-1) was observed, which overwhelmed all other vibrations (those of
bismuth oxide) expected in this region. Bands of hydroxyl groups and water (between 3700 and
2700 cm-1 and at 1625 cm-1, respectively) and a doublet of carbonate bands at 1458 and 1392
cm-1 were also observed.
In the BR composite, Bi-O vibrations of bismuth oxide are observed below 750 cm-1. Bands
characteristic of carbonates (845, 1392 and 1458 cm-1) confirm presence of (BiO)2CO3, which
was identified also by XRD examination (Fig. 4). Vibrations from C=C sheets in rGO are
confirmed by the band at 1625 cm-1. Bands at 1110 and 1197 cm-1 suggest presence of C-O-C
ether bonds, while the band at 1572 cm-1 is attributed to C=C aromatic stretching.
The FTIR spectrum of TR composite is dominated by Ti-O lattice vibrations below 1000 cm-1
and hydroxyl/water vibrations. The latter is overlapping with C=C vibration of graphene oxide.
Several bands give away presence of carbonyl (1680 and 1718 cm-1), carboxylic (1411 and
1240 cm-1) and ether (1122 and 1060 cm-1) functional groups.
Regarding the TBR composite, the FTIR spectrum is also dominated by lattice Ti-O and
water/hydroxyl vibrations. The water band at 1633 cm-1 exhibits a notable shoulder at lower
wavelengths, characteristic of C=C vibrations of reduced graphene oxide. Also, a band at 1688
cm-1 is apparent which can be attributed to surface carbonyl groups. Bands characteristic of
ether functional groups also appear at 1110 and 1197 cm-1. Bands characteristic of (BiO)2CO3
are weak, indicating this phase is poorly represented. This is in line with XRD analysis (Fig.
4), where no signal of crystalline (BiO)2CO3 could be observed.
12

Combined results of transmission FTIR and XRD analyses show that: (i) sample B contains
mainly -BiO with trace amounts of bismuth carbonate, TB composite consists mainly of
(BiO)2CO3 and traces of -BiO, BR composite consists mainly of -BiO and traces of
(BiO)2CO3, whereas TBR composite contains amorphous Bi2O3 coexisting with traces of -
BiO and (BiO)2CO3 phases; (ii) different oxygen containing functional groups are still present
in reduced graphene oxide after hydrothermal reduction: ether groups in presence of Bi2O3
and/or (BiO)2CO3 (BR composite), carbonyl groups in presence of TiO2 (TR composite), and
both ether and carbonyl groups in TBR composite.
3.1.4. UV-Vis DR analysis
Optical properties of prepared materials are presented in Fig. 6. Pure T sample shows strong
absorption in the UV light region (400 to 200 nm) with a band gap of 3.28 eV, typical for
anatase TiO2 [37,38]. On the other hand, B sample shows strong visible light absorption and a
band gap of 2.4 eV, typical for β-Bi2O3 [17]. The UV-Vis DR spectra of TB and TBR samples
show strong absorption in the 550-200 nm range, indicating that prepared samples have a strong
visible and UV light response, which is in contrast to pure T sample. There are two contributions
observed in the acquired UV-Vis DR spectra: (i) the first one between 200 and 400 nm is related
to absorption of UV light by T, and (ii) the second one above 500 nm corresponds to the
absorption of visible light by β-Bi2O3. According to the literature [28] the band gap of
(BiO)2CO3 is around 3.1-3.2 eV, meaning it can be activated by incident light with <400 nm.
As a result, (BiO)2CO3 does not act as a visible light photosensitizer, but rather to accommodate
photo-generated electrons originating from the conduction band of β-Bi2O3. The increasing
background absorbance (above 400 nm in spectra of TR and above 550 nm in spectra of BR
and TBR samples) is related to light absorption by rGO, which was also indicated by the color
change of these samples that become dark grey after introduction of rGO (Fig. S2).
3.1.5. Photoelectrochemical measurements
13

The separation of photo-generated electron-hole pairs of all prepared solids was systematically
tested by photocurrent (Fig. 7) and EIS (Fig. 8) measurements. The samples produced anodic
photocurrent when the light was switched on, but only in the case of pure B sample as well
as TB and TBR composites we can speak of a significant increase of the anodic current.
Concerning TB sample, photoelectrochemical measurements show that a heterojunction
between T phase and β-Bi2O3, as well as between β-Bi2O3 and (BiO)2CO3 is formed. β-Bi2O3
works as a visible light photosensitizer for T phase.
The increase of the photocurrent density for TBR composite (for 30 % in the eighth
illumination cycle in comparison to TB catalyst) clearly shows that rGO acts as a web for
charge carriers percolation which prolongs their separation and consequently results in a higher
photocurrent measured. High photocurrent density means that more charge carriers are
available to produce hydroxyl radicals as active oxygen species, which can participate in the
photocatalytic reaction [7,39].
The steady-state photocurrent density of pure B is approximately 20 and 30 fold lower in
comparison to TB and TBR solids, respectively, which could be attributed to the low band gap
energy of -Bi2O3 and consequently fast electron-hole recombination.
The combination of B and rGO (BR sample) did not result in increased photocurrent density,
which was also reported by other authors [40]. By looking into the CB position of rGO, it was
observed that it lies at -0.08 eV vs. NHE, [41]. Considering that the CB of -Bi2O3 lies at 0.16
eV (Fig. 10), it is too positive to enable transfer of photo-excited electrons from Bi2O3 to rGO.
As a result, no increased photocurrent density is observed when rGO and -Bi2O3 are coupled.
On the other hand, the CB of TiO2 lies at -0.29 eV, which is appropriate for the photo-generated
electrons to migrate to rGO and thus segregate efficiently.
The photocurrent densities of T and TR samples are very low, which is attributed to their band
gap energy (3.28 eV, see UV-Vis DR results, Fig. 6). The increase of photocurrent densities for
14

these two samples is attributed to the fact that although UV cut-off filter was used, not all UV
irradiation was eliminated (Fig. S1) which triggered the catalytic activity of TiO2. Photocurrent
density measurements show stable response for all materials in the last four out of eight
subsequent illumination cycles. The decrease of photocurrent densities of some catalysts (B,
BR and TB) during the first four illumination cycles is ascribed to the stabilization of the
measuring system (i.e. dynamic equilibrium at the solid-liquid interface).
To further study charge transfer processes at the contact between electrode (surface of tested
catalysts) and electrolyte (0.1 M Na2SO4), electrochemical impedance spectroscopy (EIS) was
utilized. The results are present in the form of EIS Nyquist plots in Fig. 8, where the
intermediate-frequency response is related to electron transport and transfer at the electrode
(catalyst surface) - electrolyte (Na2SO4) interface [5]. The diameter of semicircles in the
Nyquist plots reflects the charge transfer process, and the smaller diameter of semicircle
presents the lower charge transfer resistance [42]. The TB sample and especially TBR catalyst
have smaller semicircles in comparison to other tested catalysts. The semicircle diameter
belonging to the TBR sample is smallest when rGO is present, indicating that the recombination
of electron-hole pairs is further reduced after rGO addition. Nyquist plots of T and TB catalysts
support the results of photocurrent measurements (Fig. 7) showing that there is a junction
between TiO2 and β-Bi2O3, as well as between β-Bi2O3 and (BiO)2CO3, resulting in more
efficient charge carrier separation. Wide semicircle diameters of B and BR samples show that
the low band gap energies of B and BR favor the recombination of electron-hole pairs.
Based on the results of UV-Vis DR, photocurrent and EIS measurements, we can predict that
any meaningful photocatalytic activity of T based catalysts without B present can only be
triggered by illumination with UV light. On the other hand, appreciable catalytic activity under
visible light illumination is expected for TB and especially TBR composite. This postulation
was tested in the subsequent experiments of photocatalytic BPA degradation.
15

3.2. Photocatalytic BPA oxidation
Fig. 9 shows BPA degradation curves obtained in the presence of synthesized materials. In the
performed runs no leaching of catalyst constituents into the liquid phase was observed, as
confirmed by means of ICP-OES analysis. The experiments were first conducted for 30 min
without illumination (dark phase) to determine the extent of BPA adsorption on the catalyst
surface. The obtained results show that the decrease of BPA concentration in the dark phase
was below 3 %, which means than an influence of adsorption on measured data could be
neglected. After 30 min, the BPA solution containing dispersed catalyst powder was illuminated
with the same halogen lamp as used in the photocurrent measurements. The obtained c/c0 vs.
time dependencies were found to be reproducible with an error of conversions measured
within ±1 %. The following decreasing order of BPA degradation was achieved: TBR > TB
> TR > T > B > BR. This is completely in line with decreasing order of photocurrent density.
As a result, we can state that the photocurrent density corresponds exactly to the
photocatalytic activity and can be used as a reliable descriptor for predicting the
photocatalytic activity. Degradation of BPA carried out over TBR and TB composites was
significantly higher (94 and 80 % BPA conversions were achieved in 120 min, respectively)
in comparison to all other solids examined in this work. This result is not surprising if we take
into the consideration the highest photocurrent density measured under visible light
illumination by TBR sample (Fig. 7), meaning that in the presence of this material more
charge carriers are available to generate reactive OH· radicals which participate in
photocatalytic oxidation reaction. Degradation of BPA when using TB composite was
significantly higher in comparison to TR sample (34 % BPA conversion). The results show
that in the TB composite a junction between T, β-Bi2O3 and (BiO)2CO3 was formed and that
β-Bi2O3 is acting as a photosensitizer for T and (BiO)2CO3 under visible light illumination.
16

Combining T and rGO is an efficient way to boost photocatalytic activity of TiO2 under UV
light illumination, as rGO acts as a sink for electrons which prolongs their lifetime before
recombination with holes [4]. Indeed, in our test we measured higher photocatalytic activity
of TR sample compared to that of pure T (BPA conversions of 34 and 26 %, respectively).
Pure T exhibited some photocatalytic activity for BPA removal despite the fact that UV-Vis
DR and photocurrent measurements suggest its photocatalytic activity under visible light as
negligible. The observed behavior is likely triggered by some minority surface defect sites [43],
which are able to absorb visible light and generate reactive OH radicals that ultimately
participate in the catalytic reaction.
Although pure B sample exhibits the highest absorption of visible light among all tested samples
(Fig. 6), its low BET specific surface area, especially narrow band gap and inappropriate
position of the conduction band promote recombination of electron-hole pairs which in turn
ultimately hinders its photocatalytic activity. The BR composite shows lower photocatalytic
activity compared to pure -Bi2O3, very likely as a result of its partial transformation to
(BiO)2CO3. The conduction band of (BiO)2CO3 is 3.1 eV [28]; it is too wide to allow visible
light absorption, thus making this material inappropriate as a visible light driven photocatalyst.
At the end of each BPA degradation run, TOC value of the reaction solution was measured in
order to evaluate the extent of BPA mineralization. The results of TOC measurements (Table
2) are in good correlation with the obtained BPA conversions (Fig. 9): the highest degree of
BPA mineralization was achieved over TB and TBR composites. CHNS elemental analysis was
further performed on fresh and spent catalyst samples (Table 2) to determine the true TOC
removal (TOCm) and TOCa values corresponding to carbon-containing species accumulated on
the catalyst surface, which constitute the TOCR value. The listed results reveal that the amount
of carbon deposited on spent samples was higher than on fresh solids, which is a sign that BPA
and/or partially oxidized BPA degradation products adsorb on the catalyst surface. This can be
17

most probably attributed to high BET specific surface area of T sample (or T component in TB
composite).
The BPA degradation rates in the second half of performed oxidation runs (Fig. 9, t > 60 min)
conducted in the presence of B, T, TR and BR catalysts are zero (d(C/C0,BPA)/dt=0), which is a
sign that these catalysts were deactivated by adsorption of partially oxidized reaction
intermediates onto the catalyst surface that block active sites. CHNS elemental analysis results
of fresh and spent B catalyst also show that TOCa is higher than TOCm (1.4 and 2.6 %,
respectively, Table 2), revealing that the predominant pathway of BPA removal from the liquid
phase over this catalyst is actually adsorption of partially oxidized reaction intermediates. The
same is true for T sample, as TOCa and TOCm values of 4 and 7 % were calculated.
In the case of TB composite, the results of CHNS elemental analysis clearly show that the main
pathway of BPA removal is photocatalytic oxidation (43 % TOCm) and not adsorption of
partially oxidized reaction intermediates (8 % TOCa). Composites containing rGO were not
subjected to CHNS elemental analysis, because rGO influences the measurements which could
lead to false conclusions.
We performed FTIR and XRD analyses of spent TB and TBR catalysts. XRD results (Fig. S3)
showed that no changes of the crystalline components (TiO2, -Bi2O3 and (BiO)2CO3) occur
during photocatalytic runs, indicating that all these crystalline phases are stable under the
employed reaction conditions. FTIR results showed an increase of band intensity at 1572 and
1688 cm-1 in spent TBR catalyst. These vibrations occur due to the presence of C=C bonds in
aromatic compounds as well as of carbonyl group containing species, respectively, indicating
that carbonaceous deposits were formed during photocatalytic oxidation of BPA, or during
oxidation of the rGO component. These deposits are not observed in spent TB catalyst, where
the rGO component is absent. We tested also the reuse of TBR catalyst for BPA degradation.
The results of the second run (Fig. S5) show that conversion of BPA was 5 % lower in
18

comparison to the first run (94 and 89 %, respectively). The TOC removal was also lower (60
vs. 50 % for the first and second run, respectively). The decrease of TBR catalyst activity was
noted in further oxidation cycles (Fig. S5). The XRD and FTIR measurements performed on
spent TBR sample show that the catalyst is structurally stable under employed reaction
conditions, so deactivation is likely related to adsorption of organics.
3.3. Proposed charge carrier migration cascade
The valence band edge (EVB) and the conduction band edge (ECB) of T and B catalysts were
calculated accordingly to the Mulliken electronegativity theory [44,45]:
퐸푉퐵 = 푋 − 퐸푒 + 0.5 ∗ 퐸퐵퐺
(1)
where EVB is the VB edge potential, X is the electronegativity of the semiconductor, Ee is the
energy of free electrons on the hydrogen scale (about 4.5 eV) and EBG is the corresponding band
gap energy of the semiconductor. The conduction band edge can be calculated as follows:
퐸퐶퐵 = 퐸푉퐵 − 퐸퐵퐺
(2)
On the basis of UV-Vis DR measurements (Fig. 6), the EBG values of T and B catalysts are
equal to 3.2 and 2.45 eV, respectively. Accordingly to Xu and Schoonen [45], their
corresponding electronegativities are 5.81 and 6.21 eV, respectively. Consequently, calculated
EVB and ECB are 2.91 and -0.29 eV for pure T, and 2.93 and 0.48 eV for B catalyst. The low
conduction band edge of B catalyst (0.48 eV) cannot provide sufficiently negative potential for
o
-
∙
the photo-excited electrons to scavenge the adsorbed O2 (E (O2/O2 )=-0.33 V vs. NHE and
o
∙
E (O2/HO2 )=-0.05 V vs. NHE) [20,46]. This in turn causes fast recombination of photo-excited
electron-hole pairs. This property of B catalyst was well displayed by the results of photocurrent
and EIS measurements (Figs. 7 and 8) and photocatalytic BPA degradation tests (Fig. 9).
Based on the above presented results and discussion, a plausible charge carrier migration
cascade for TB composite catalyst is illustrated in Fig. 10a. In this material, a heterojunction
between TiO2 (T) and β-Bi2O3 is formed which supports the “transfer” of photo-generated holes
19

(h+) from VB of β-Bi2O3 to upper lying VB of T. The holes can react with hydroxide ions to
form hydroxyl radicals (reaction 3), which actively participate in the degradation of water
dissolved BPA molecules:
푂퐻− + ℎ+ → 푂퐻∙
(3)
The electrons (e-) in the CB of -Bi2O3 can react accordingly to reactions 4 and 5, which are
pH dependent [47-]:
푂2 + 2퐻+ + 2푒− → 퐻2푂2
푂2 + 4퐻+ + 4푒− → 2퐻2푂
퐸푂 = +0.682 푉 (푣푠 푁퐻퐸)
퐸푂 = +1.23 푉 (푣푠 푁퐻퐸)
(4)
(5)
The VB edge of β-Bi2O3 is positioned lower than the VB edge of TiO2 [50], thus making the
transfer of h+ in this direction thermodynamically possible. Photo-generated h+ in the VB of β-
Bi2O3 as well as those transferred into the VB of T can react with H2O or OH- adsorbed on the
surface of both components.
On the other hand, recombination of photo-generated electron-hole pairs can be effectively
suppressed when a p-n junction is formed between tightly chemically bonded n-type
semiconductor TiO2 and p-type semiconductor Bi2O3 [51]. Before contact, the Fermi level of
p-type -Bi2O3 is more negative than that of n-type TiO2 due to its smaller work function
[52,53]. When a chemical contact is established between TiO2 and Bi2O3, the Fermi level of -
Bi2O3 is moved up and the Fermi level of TiO2 is moved down until an equilibrium state of
Fermi levels (EF) and inner electric field is formed at the interface between TiO2 and -Bi2O3
[54,55]. As shown in Fig. 10b, photo-generated e- at the CB of -Bi2O3 can migrate to TiO2,
since only -Bi2O3 can produce photo-generated charge carriers under visible light irradiation.
Both XRD and FTIR analyses confirmed (BiO)2CO3 in the TB catalyst. Accordingly to the
literature, the band gap of (BiO)2CO3 is around 3.1 eV [28] and the values of EVB and ECB
calculated by means of Eqs. 2 and 3 are 0.16 and 3.32 eV, respectively. The energy band
20

structures of β-Bi2O3 and (BiO)2CO3 which are not in a contact, are illustrated in Fig. 10a.
Comparison of the band positions shows that in this case the band structure is unfavorable for
the separation of photo-generated electron-hole pairs. However, when a contact between β-
Bi2O3 and (BiO)2CO3 is established, a p-n junction is formed since β-Bi2O3 is a p-type
semiconductor, and (BiO)2CO3 is known as a n-type semiconductor. Because of the p-n junction
the Fermi levels (EF) of semiconductors will reach an equilibrium state and a newly formed
energy band structure is formed as shown in Fig. 10c. It is necessary to point out that the
adjusted energy band structure is favorable for the separation of visible light generated electrons
from β-Bi2O3 to (BiO)2CO3 due to the effect of inner electric field analogous to other kinds of
p-n junctions [25,27,28]. As demonstrated before (Figs. 7 and 9), the improved charge carrier
separation in TB composite resulted in higher photocurrent during electrochemical
measurements and enhanced activity in the process of photocatalytic BPA oxidation. To
conclude, the role of β-Bi2O3 in TB composite is to act as a visible light photosensitizer as the
band gap energies of both T and (BiO)2CO3 are inappropriate for visible light absorption.
In the TBR composite, the photo-generated charge carriers in TB are injected to the rGO that
acts as a web for their percolation. The migration of charge carriers is significantly improved,
meaning that also the recombination of electron-hole pairs is additionally hindered when
compared to the TB composite (see, for instance, results of photocurrent measurements
displayed in Fig. 7 and EIS measurements depicted in Fig. 8). The extended lifetime of photo-
generated charge carriers and consequently enhanced formation of reactive radical species (OH·
and O2· can be formed from water on TiO2 via the one hole reaction, or on -Bi2O3 and
(BiO)2CO3 via slower two electron reactions, see Eqs. 4 and 5) due to percolation by rGO is
well expressed by the results of photocatalytic BPA degradation tests (Fig. 9), where the highest
degradation and mineralization of BPA were achieved in the presence of TBR composite.
4. Conclusions
21

Several composites containing TiO2, bismuth oxide and bismuth carbonate as well as rGO were
synthesized, characterized and tested for photocatalytic activity in BPA degradation. The
morphology of TiO2 nanorods was not influenced by the synthesis conditions employed in the
subsequent preparation procedure of TB and TBR composites. However, -Bi2O3 was
transformed into (BiO)2CO3 at different extents. In all bismuth containing composites, light
absorption was shifted into the visible range indicating its appropriateness to act as a
photosensitizer for TiO2. In the TB composite, a heterojunction between β-Bi2O3 and TiO2 is
formed, which allows the transfer of visible light photo-generated h+ from the VB of β-Bi2O3
to upper lying VB of TiO2. In addition, a p-n junction between TiO2 and β-Bi2O3 can be formed
enabling the transfer of photo-generated e- in CB of β-Bi2O3 to CB of TiO2. Also, (BiO)2CO3
that appears in the TB composite forms a p-n junction with β-Bi2O3, which enables the transfer
of photo-generated e- in β-Bi2O3 to (BiO)2CO3. Consequently, recombination of electron-hole
pairs in the TB composite is significantly hindered, which was confirmed by
photoelectrochemical measurements. The photocurrent triggered by visible light in B and BR
composite catalysts was low, which is attributed to the narrow band gap of β-Bi2O3 favoring
the recombination of electron-hole pairs. The function of rGO in TBR composite catalyst is to
act as a web for the percolation of charge carriers which prolongs their lifetime, making more
charge carriers available for utilization in catalytic reactions, as compared to TB composite.
This was confirmed by photoelectrochemical measurements, as 4 wt. % rGO content evidently
improved the photocurrent density. The TBR composite exhibited the highest extent of BPA
degradation and mineralization among all tested catalysts. XRD analysis of the spent TBR
catalyst showed that it is structurally stable, hence deactivation observed during the catalyst
reuse was likely caused by accumulation of carbonaceous species and/or oxidation of rGO.
Acknowledgements - The authors gratefully acknowledge the Slovenian Research Agency
(ARRS) for financial support through Research Program No. P2-0150. Mr. Janvit Teržan and
22

Mr. Kristijan Lorber are kindly acknowledged for help with transmission FTIR measurements
and photocatalytic tests.
References
1 K. Hashimoto, H. Irie, A. Fujishima, Jpn. J. Appl. Phys. 12 (2005) 8269-8285.
2 M. Ge, C. Cao, J. Huang, S. Li, Z. Chen, K.-Q. Zhang, S. S. Al-Deyab, Y. Lai, J. Mater.
Chem. A 4 (2016) 6772-6801.
3 J. Zhang, Q. Xu, Z. Feng, M. Li, C. Li, Angew. Chem. 9 (2008) 1766-1769.
4 G. Žerjav, M.S. Arshad, P. Djinović, I. Junkar, J. Kovač, J. Zavašnik, A. Pintar, Nanoscale,
9 (2017) 4578-4592.
5 F.-X. Xiao, J. Miao, B. Liu, J. Am. Chem. Soc. 136 (2014) 1559-1569.
6 A. Adán-Más, D. Wei, Nanomaterials, 3 (2013) 325-356.
7 P. Dong, Y. Wang, L. Guo, B. Liu, S. Xin, J. Zhang, Y. Shi, W. Zeng, S. Yin, Nanoscale 4
(2012) 4641-4649.
8 G.P. Sing, K.M. Shrestha, A. Nepal, K.J. Klabunde, C.M. Sorensen, Nanotechology 26
(2014) 265701-265712.
9 Y. Liu, RSC Adv. 4 (2014) 36040-36045.
10 N. Zhang, M.-Q. Yang, S. Liu, Y. Sun, Y.-J. Xu, Chem. Rev. 18 (2015) 10307-10377.
11 M.-Q. Yang, N. Zhang, M. Pagliaro, Y.-J. Xu, Chem. Soc. Rev. 43 (2014) 8240-8254.
12 N. Zhang,Y. Zhang, Y.-J. Xu, Nanoscale 4 (2012) 5792-5813.
13 S. Malato, P. Fernandez-Ibanez, M.I. Maldonato, M. Blanco, W. Gernjak, Catal. Today,
147 (2009) 1-59.
14 X. Zhang, L.Z. Zhang, T.F. Xie, D.J. Wang, J. Phys. Chem. C 113 (2009) 7371-7378.
15 F. Duo, Y. Wang, C. Fan, X. Mao, X. Zhang, Y. Wang, J. Liu, Mater. Charact. 99 (2015)
8-16.
16 L. Zhang, G. Tann, S. Wei, H. Ren, A. Xia, Y. Luo, Ceram. Int. 39 (2013) 8597-8604.
17 H. Cheng, B. Huang, J. Lu, Z. Wang, B. Xu, X. Qin, X. Zhang, Y. Dai, Phys. Chem.
Chem. Phys. 12 (2010) 15468-15475.
18 H.W. Kim, Thin Solid Films 516 (2008) 3665-3668.
19 J. Hou, C. Yang, Z. Wang, W. Zhou, S. Jiao, H. Zhu, Appl. Catal. B 142-143 (2013) 504-
511.
20 H.Y. Jiang, K. Cheng, J. Lin, Phys. Chem. Chem. Phys. 14 (2012) 12114-12121.
23

21 A. Hameed, T. Montini, V. Gombac, P. Fornasiero, J. Am. Chem. Soc. 130 (2008) 9658-
9659.
22 L. Li, X. Huang, T. Hu, J. Wang, W. Zhang, J. Zhang, New. J. Chem 38 (2014) 5293-
5302.
23 P. Wang, L. Xu, Y. Ao, C. Wang, J. Colloid. Interf. Sci. 495 (2017) 122-129.
24 Y. Bian, Y. Ma, Y, Shang, P. Tan, J. Pan, Appl. Surf. Sci. (2017)
https://doi.org/10.1016/j.apsusc.2017.06.063.
25 G. Cai, L. Xu, B. Wei, J. Che, H. Gao, W. Sun, Mater. Lett. 120 (2014) 1-4.
26 Y. Huang, W. Fan, B. Long, H. Li, F. Zhao, Z. Liu, Y. Tong, H. Ji, Appl. Catal. B 185
(2016) 68-76.
27 R. Hu, X. Xiao, S. Tu, X. Zuo, J. Nan, Appl. Catal. B 163 (2015) 510-519.
28 G. Zhu, Y. Liu, M. Hojamberdiev, J. Han, J. Rodríguez, S. A. Bilmes, P. Liu, New J.
Chem. 39 (2015) 9557-9568.
29 C.A. Staples, P.B. Dome, G.M. Klecka, S.T. Oblock, L.R. Harris, Chemosphere 36 (1998)
2149-2173.
30 B. Erjavec, T. Tišler, R. Kaplan, A. Pintar, Ind. Eng. Chem. Res. 52 (2013) 12559-12566.
31 G.P. Sing, K.M. Shrestha, A. Nepal, K.J. Klabunde, C.M. Sorensen, Nanotechnology, 26
(2014) 265701-265712.
32 B. Erjavec, R. Kaplan, A. Pintar, Catal. Today 241 (2015) 15-24.
33 P.T.N. Nguyen, C. Salim, W. Kurniawan, H. Hinode, Catal. Today 230 (2014) 166-173.
34 D. Liang, C. Cui, H. Hu, Y. Wang, S. Xu, B. Ying, P. Li, B. Lu, H. Shen, J. Alloy. Compd.
582 (2014) 236-240.
35 L. Liu, J. Jiang, S. Jin, Z. Xia, M. Tang, CrystEngComm 13 (2011) 2529-2532.
36 J.Z. Marinho, L.M. Santos, L.R. Macario, E. Longo, A.E.H. Machado, A.O.T. Patrocinio,
R.C. Lima, J. Braz. Chem. Soc. 26 (2015) 498-505.
37 J. Zhang, P. Zhou, J. Liu, J. Yu, Phys. Chem. Chem. Phys. 16 (2014) 20382-20386.
38 C. Dette, M.A. Pérez-Osorio, C.S. Kley, P. Punke, C.E. Patrick, P. Jacobson, F. Giustino,
S.J. Jung, K. Kern, Nano Lett. 14 (2014) 6533-6538.
39 A. Adán-Más, D. Wei, Nanomaterials 3 (2013) 325-356.
40 X. Liu, L. Pan, T. Lv, Z. Sun, C.Q. Sun, J. Colloid Interface Sci. 408 (2013) 145-150.
41 S. Singha, N. Khare, RSC Adv. 5 (2015) 96562-96572.
42 X. T. Wang, C. H. Liow, D. P. Qi, B. W. Zhu,W. R. Leow, H. Wang, C. Xue, X. D. Chen,
S. Z. Li, Adv.Mater. 26 (2014) 3506–3512.
43 W. Göpel, A. Anderson, D. Frankel, M. Jaehnig, K. Phillips, J.A. Schäfer, G. Rocker, Surf.
Sci. 139 (1984) 333-346.
24

44 M.A. Butler, D.S. Ginley, J. Electrochem. Soc. 125 (1987) 228-232.
45 Y. Xu, M.A.A. Schoonen, Am. Mineral. 85 (2000) 543-556.
46 D.T. Sawyer, J.S. Valentine, Acc. Chem. Res. 14 (1981) 393-400.
47 R. Abe, H. Takami, N. Murani, B. Ohtani, J. Am. Chem. Soc. 130 2008 7780-7781.
48 G. Liu, L. Wan, C. Sun, X. Wang, Z. Chen, S.C. Smith, H.M. Cheng, G.Q. Lu, Chem.
Mater. 21 (2009) 1266-1274.
49 P.Y. Ayekoe, D. Robert, D.L. Gone, Environ. Chem. Lett. 14 (2016) 387-393.
50 D. Beydoun, R. Amal, G. Low, S. McEvoy, J. Mol. Catal. A 180 (2002) 193-200.
51 Y. Huang, Y. Wei, J. Wang, D. Luo, L. Fan, J. Wu, Appl. Surf. Sci. 423 (2017) 119-130.
52 S. Yi, X. Yue, D. Xu, Z. Liu, F. Zhao, D. Wangab, Y. Lin, New J. Chem. 39 (2015) 2917-
2924.
53 N. Wei, H. Cui, C. Wang, G. Zhang, Q. Song, W. Sun, X. Song, M. Sun, J. Tian, J. Am.
Ceram. Soc. 100 (2017) 1339-1349.
54 Z. Zhang, C. Shao, X. Li, C. Wang, M. Zhang, Y. Liu, ACS Appl. Mater. Interfaces 2
(2010) 2915-2923.
55 G. Dai, J. Yu, G. Liu, J. Phys. Chem. C 115 (2011) 7339–7346.
25

Fig. 1. SEM images of prepared catalysts.
26

0.035
0.030
0.025
0.020
0.015
0.010
0.005
0.000
B
T
TR
BR
TR
TBR
1
10
100
Pore diameter, nm
Fig. 2. a) N2 adsorption-desorption isotherms and b) BJH pore size distribution of synthesized
catalysts.
27

Fig. 3. SEM-EDX mapping of TBR sample.
28

GO
BR
T
B
TR
TB
TBR
5
15
25
35
45
55
65
2
Fig. 4. XRD patterns of prepared single-phase catalysts as well as multi-phase composites.
Vertical blue lines indicate most intensive reflections of (BiO)2CO3 (JCPDS 00-041-1488),
while red and green lines indicate reflections of anatase TiO2 (JCPDS 00-021-1272) and -
Bi2O3 (JCPDS 00-027-0050), respectively.
29