Preparation by sonophotodeposition method of bimetallic photocatalysts Pd-Cu/TiO2 for sustainable gaseous selective oxidation of methanol to methyl formate

Article abstract of DOI:10.1016/j.molcata.2015.10.031

It has been demonstrated that sonophotodeposition can be one choice as a green method to synthesize bimetallic supported photocatalysts with enhanced performance for selective oxidations. A series of Pd-Cu supported on Titania-P90 photocatalysts were successfully prepared using this innovative method of effective synergistic combination of sonication and light. In addition, our method does not require the use of strong chemical reduction agent and it is executed in a short time, room temperature and atmospheric pressure. The prepared materials were characterized by a number of techniques such as High-Resolution Transmission Electron Microscopy (HRTEM), DR UV-vis spectroscopy, X-ray Photoelectron Spectroscopy (XPS) and powder X-ray diffraction (XRD). Additionally, better bimetallic systems were obtained (methanol conversion >50% and selectivity to methyl formate >80%) by SonoPhotoDeposition (SPD) than in the case of the conventional photodeposition methodology. It has been discussed the possible reasons of the observed slight deactivation (8% after 2 h reaction test) of the best performing material in gas phase methanol selective photo-oxidation.

Full text of DOI:10.1016/j.molcata.2015.10.031

Journal of Molecular Catalysis A: Chemical 411 (2016) 247–256
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Preparation by sonophotodeposition method of bimetallic
photocatalysts Pd–Cu/TiO2 for sustainable gaseous selective oxidation
of methanol to methyl formate
Paweł Lisowski∗, Juan C. Colmenares∗, Dariusz Łomot, Olga Chernyayeva,
Dmytro Lisovytskiy
Institute of Physical Chemistry PAS, Kasprzaka 44/52, 01-224 Warsaw, Poland
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 5 October 2015
Received in revised form 27 October 2015
Accepted 29 October 2015
Available online 2 November 2015
It has been demonstrated that sonophotodeposition can be one choice as a green method to synthe-
size bimetallic supported photocatalysts with enhanced performance for selective oxidations. A series of
Pd–Cu supported on Titania-P90 photocatalysts were successfully prepared using this innovative method
of effective synergistic combination of sonication and light. In addition, our method does not require
the use of strong chemical reduction agent and it is executed in a short time, room temperature and
atmospheric pressure. The prepared materials were characterized by a number of techniques such as
High-Resolution Transmission Electron Microscopy (HRTEM), DR UV–vis spectroscopy, X-ray Photoelec-
tron Spectroscopy (XPS) and powder X-ray diffraction (XRD). Additionally, better bimetallic systems were
obtained (methanol conversion >50% and selectivity to methyl formate >80%) by SonoPhotoDeposition
(SPD) than in the case of the conventional photodeposition methodology. It has been discussed the possi-
ble reasons of the observed slight deactivation (8% after 2 h reaction test) of the best performing material
in gas phase methanol selective photo-oxidation.
Keywords:
Sonophotodeposition
Bimetallic
TiO2
Sonication
Methanol photo-oxidation
© 2015 Elsevier B.V. All rights reserved.
1. Introduction
sibilities for the synthesis of a broad spectrum of nanostructured
materials [5]. Ultrasonic treatment is one of the emerging tools that
Photocatalysts for selective oxidation must have, besides an
appropriate location of the valence and conduction bands, a good
absorption of photons, a long lifetime of the photoactivated species,
a good adsorption of reactants and a relatively easy desorption of
products. In this sense, structure, particle size and surface char-
acteristics have been found to influence not only the activity, but
also the selectivity of photocatalysed reactions [1–2]. One of the
main advantages of using photocatalysis is the ability to operate
reactions at room temperature, and this is especially relevant for
selective oxidation reactions. Selective oxidation using photocatal-
ysis potentially offers an alternative, safer and greener route for the
synthesis of valuable chemicals [3–4].
could be the alternative to thermal processing. It promotes the reac-
tion under milder conditions where drastic conditions are required
conventionally. Ultrasound functions by acoustic cavitation process
that involves sequential formation, growth and collapse of micro-
scopic vapor bubbles in the liquid. These localized hot spots have
temperatures of roughly 5000 ◦C, pressures of about 500 atmo-
spheres, and lifetimes of a few microseconds [6–7]. Suslick [8]
performed much of the early work exploring the effects of acoustic
cavitation on materials synthesis. The conditions produced during
ultrasonic irradiation can be described as unusual, comparing to
traditional energy sources, and they cannot be realized by other
methods.
In recent years, the development of novel environmental
friendly and cost efficient methods for materials preparation that
could replace the old ones is on demand. Unconventional and “soft”
techniques such as sonication and photochemistry offer huge pos-
Photocatalyst deactivation phenomena are more predominant
in gas phase than in aqueous phase, because in the latter the
water molecules restore the surface hydroxylation and assist in
the removal of the adsorbed species from the photocatalyst sur-
face [9]. In gas phase reactions, intermediate species with slower
kinetics and higher adsorption affinity than the target pollutant
cause a deactivation that may sometimes be reversed by favouring
their desorption or their photocatalytic degradation. Deactivation
in the gas phase can differ in nature, and it depends on the type
∗
Corresponding author.
E-mail addresses: plisowski@ichf.edu.pl (P. Lisowski),
jcarloscolmenares@ichf.edu.pl (J.C. Colmenares).
http://dx.doi.org/10.1016/j.molcata.2015.10.031
1381-1169/© 2015 Elsevier B.V. All rights reserved.

248
P. Lisowski et al. / Journal of Molecular Catalysis A: Chemical 411 (2016) 247–256
of organic substrate being oxidized and also on the conditions
of the photoprocess. An analysis of the open literature reveals
that photocatalyst deactivation is generally found in continuous-
flow photocatalytic reactors with a surface attached catalyst [10].
The catalyst may be deactivated either by formation of surface
intermediates with higher adsorption ability than the target pol-
lutant (reversible deactivation) or by sticky “heavy” products that
are difficult to decompose or desorb (irreversible deactivation).
Peral and Ollis [11] noted catalyst deactivation when photo-
oxidizing 1-butanol and butyraldehyde in batch reactor. These
authors proposed that the formation of strongly adsorbed butanoic
acid in both cases was responsible for the catalyst deactivation.
The deactivation of TiO2 during the gas-phase photooxidation of
trichloroethylene was reported by Larson and Falconer [12]. These
researchers indicated that apparently strongly bound species, such
as carbonates, accumulated on the surface deactivating the catalyst.
Vorontsov et al. [13] investigated temperature deactivation of TiO2
for acetone oxidation, and proposed that the accumulated surface
products resulting from thermal oxidation of acetone caused deac-
tivation. Méndez-Román and Cardona-Martıinez [14] studied the
relationship between the formation of surface species and catalyst
deactivation during the gas phase oxidation of toluene. Catalytic
deactivation caused by byproducts was observed in the photocat-
alytic conversion of triethylamine over TiO2 [15]. As pointed out by
Sauer and Ollis [10], every single-pass catalytic process will even-
tually lead to the deactivation of the catalyst, often not observed
in practice due to low levels of substrate or experiments carried
out using short periods of time, or both. A better understanding of
deactivation processes is essential for improving and optimizing
process conditions, the catalysts themselves and for circumvent-
ing premature catalyst degradation in order to minimize additional
costs.
Our systematic study of the photocatalytic oxidation of volatile
organic compounds have led to receive highly active and selective
photocatalysts prepared via SonoPhotodeposition (SPD) advanced
methodology [16]. The advantages of SPD methodology over con-
ventional methods include: preparation at room temperature and
atmospheric pressure, no need to use reducing agents and very
short reaction times. Furthermore, the physicochemical proper-
ties (high surface area and phase purity, particles with different
sizes and shapes, uniform coating of nanoparticles on substrates,
and many others) of the produced photocatalytic materials can be
easily tuned by properly adjusting the parameters and conditions
adopted in their preparation. In the present study, we modified
the surface of commercial (Evonik) P90 TiO2 with bimetallic Pd–Cu
nanoparticles using SonoPhotoDeposition (SPD) advanced method-
ology. The modified Pd–Cu/TiO2 photocatalysts were studied in the
photocatalytic oxidation of methanol to methyl formate under UV
illumination. The photocatalytic activity and selectivity of these
materials were evaluated in our system for gas phase photocatalytic
oxidation of alcohols [17]. Additionally, we provided some key
insights in understanding the catalyst tendency to deactivate which
we believe is due to the accumulation of “organic residues” on
the surface of photocatalysts poisoning the Strong-Metal-Support-
Interaction (SMSI) observed for the best performing photocatalytic
system. To our best knowledge, this is the first time that a compre-
hensive study on this kind of systems has been reported.
as support. The detailed procedure was as follows: 0.1 g of oxalic
acid and desired amount of precursors of palladium and copper
were dissolved in 120 mL of H2O:CH3CN (30:70, v/v) and 0.5 g of
TiO2 was dispersed into this solution and pH was adjusted to ∼2.
The nominal palladium loading for all the bimetallic photocata-
lysts was designed as 1.0 wt.% (0.05 mmol) with different atom
content Pd/Cu ratios of 9:1, 3:1 and 1:1. The batch photoreactor
with such prepared mixture was placed into the ultrasonic bath
(35 kHz, 560 W, Sonorex Digitec-RC, Bandelin) (Fig. 1). The sus-
pension was first kept in the dark for 30 min to reach complete
adsorption equilibrium. Sonophotodeposition was performed by
illuminating the suspension for 60 min with a low pressure mer-
cury lamp (6 W, ꢀmax = 254 nm) and with ultrasonic bath switched
on. The average luminous intensity (∼0.005 W/m2) was determined
by a radiometer Model HD 2302 (supplied by DELTA OHM, Italy)
with UV-C laser power probe (220–280 nm). The synthesis reac-
tion was carried out under argon flow (flow rate 70 mL min−1) and
thermostated at 20 ◦C. Then the product was recovered by slowly
evaporation in rotary evaporator, dried at 110 ◦C for 10 h, and cal-
cined at 300 ◦C for 4 h under air flow (flow rate 30 mL min−1).
The photocatalysts were labelled as 1 wt.%. Pd–Cu(9-1)/TiO2 P90,
1 wt.%. Pd–Cu(3-1)/TiO2 P90 and 1 wt.%. Pd–Cu(1-1)/TiO2 P90.
For comparative purposes, 1 wt.%. Pd/TiO2 P90 and Cu(1)/TiO2
P90 were prepared by sonophotodeposition method, and 1 wt.%.
Pd–Cu(1-1)/TiO2 P90 was synthesized by photodeposition (with-
out ultrasounds). Unmodified TiO2 P90 was chosen as reference
material, without the addition of metal precursors.
2.2. Characterization methods
The specific surface area, pore volume, and average pore diam-
eter were determined by N2 physisorption using a Micromeritics
ASAP 2020 automated system and the Brunauer–Emmet–Teller
(BET)[18] and the Barret–Joyner–Halenda (BJH) methods [19]. Each
photocatalyst was degassed under vacuum at <1 × 10−5 bar in the
Micromeritics system at 300 ◦C for 4 h prior to N2 physisorption.
Powder XRD measurements were performed using standard
Bragg–Brentano configuration. This type of arrangement was pro-
vided using Siemens D5000 diffractometer (equipped with a
horizontal goniometer) with ꢁ−2ꢁ geometry and Ni filtered Cu K␣
radiation, powered at 40 kV and 40 mA. Data were collected in the
range of 2ꢁ = 10–90◦ with step interval of 0.02◦ and counting time
up to 5 s per step.
The average crystallite size (D in nm) was determined according
to the Scherrer equation [20]:
kꢀ
D =
(1)
ˇcosꢁ
where D is the average crystallite size of the catalyst (nm), ꢀ is the
wavelength of the Cu kꢀ X-ray radiation (ꢀ = 0.154056 nm), k is a
coefficient usually taken as 0.94, ˇ is the full width at half maxi-
mum (FWHM) intensity of the peak observed at 2ꢁ (radian), and
ꢁ is the diffraction angle. The phase contents of the samples can
be estimated from the respective XRD peak intensities using the
following equation [21]:
1
fA
=
(2)
I
1
R
A
1 +
K I
2. Experimental
K = 0.79f A > 0.2
2.1. Photocatalysts preparation
Palladium (II) acetylacetonate (35% Pd, Acros), Copper(II) acety-
lacetonate (98% Cu, Acros) and commercial TiO2 (AEROXIDE TiO2
P90, Evonik Industries) were used as Pd, Cu precursors and TiO2
K = 0.68f A ≤ 0.2
where:

P. Lisowski et al. / Journal of Molecular Catalysis A: Chemical 411 (2016) 247–256
249
Fig. 1. SonoPhotodeposition advanced methodology for mono and bimetallic photocatalysts preparation.
fA is the fraction of anatase phase in the powder, and IA and
IR are the X-ray intensities of the anatase (1 0 1) and rutile (1 1 0)
diffraction peaks, respectively.
absorbed light E. Regarding absorption threshold, it was deter-
mined according to the formula [23]:
HRTEM measurements were carried out using FEI TITAN Cubed
electron microscope operated at an acceleration voltage of 300 keV
and equipped with an energy dispersive X-ray (EDS) EDAX spec-
trometer. The samples were prepared by dispersing in pure alcohol
using ultrasonic cleaner and putting a drop of this suspension on
carbon films on copper grids and purified with plasma cleaner.
The XPS measurements were performed using a VG Scientific
photoelectron spectrometer ESCALAB-210 using Al K␣ radiation
(1486.6 eV) from an X-ray source operating at 15 kV and 20 mA.
Survey spectra were recorded for all the samples in the energy
range from 0 to 1350 eV with 0.4 eV step. High resolution spectra
were recorded with 0.1 eV step, 100 ms dwell time and 25 eV pass
energy. Sixty degrees take-off angle was used in all measurements.
The curve fitting was performed using the AVANTAGE software
provided by Thermo Electron, which describes each component of
the complex envelope as a Gaussian–Lorentzian sum function; a
constant 0.3( 0.05) G/L ratio was used. The background was fit-
ted using nonlinear Shirley model. Scofield sensivity factorsand
measured transmission function were used for quantification. Aro-
matic carbon C 1s peak at 284.5 eV was used as reference of binding
energy.
1240
Egap
ꢀ =
(4)
2.3. Photocatalytic activity measurements
The schematic representation of gas phase methanol photoox-
idation setup is given in Scheme 1 [17]. Methanol was introduced
by bubbling the air (25 cm3 min−1) through a glass saturator filled
with methanol. The saturator was immersed in a thermostat kept
at 0 ◦C. The gas flow rates were measured and controlled by mass
flow controllers (supplied by Bronkhorst HI-TEC). The flow-type
photoreactor was vertically enclosed by an aluminum foil cylin-
drical reflector (20 cm × 13 cm × 1 mm) to exclude any external
light source and maximize light energy usage within the reactor.
The photocatalyst bed height was 5 cm. The adsorption equilib-
rium reagent-photocatalyst was achieved in the dark after 2 h.
The light source was a medium pressure 125 W mercury lamp
(ꢀmax = 365 nm; supplied by Photochemical Reactors Ltd. Model
RQ3010) built into a lamp housing and centered vertically in the
reflector (2.5 cm between the lamp and photoreactor) and ther-
mostated at 30 ◦C. The average luminous intensity (∼260 mW/cm2)
was determined by a radiometer ILT 1400 (supplied by Inter-
national Light Technologies, Inc., USA) with UV–vis laser power
probe (250–675 nm). Reaction products were quantitatively ana-
lyzed by on-line gas chromatography (HP 5890 series II Hewlett
Packard-USA equipped with a flame ionization detector (FID) and
a methanizer model 510) (supplied by SRI INSTRUMENTS) and
identified by gas chromatography coupled with mass spectrom-
etry (HP-5 column GC (6890 Series)–MS(5973) Hewlett Packard
equipped with FID and TCD Detectors).
UV–vis Diffuse Reflectance spectroscopy was performed using
a UV/VIS/NIR spectrophotometer Jasco V-570 equipped with an
integrating sphere. The baseline was recorded using SpectralonTM
(poly(tetrafluoroethylene)) as a reference material. Band-gaps val-
ues were calculated based on the Kubelka–Munk functions [22] f(R),
which are proportional to the absorption of radiation, by plotting
[f(R) h␯)]1/2 against h␯. The function f(R) was calculated using Eq.
(3):
(1 − R)2
f
R =
( )
(3)
2R
Band gap values were obtained from the plot of the
Kubelka–Munk function [F(R )E]1/2 versus the energy of the
∞

250
P. Lisowski et al. / Journal of Molecular Catalysis A: Chemical 411 (2016) 247–256
Scheme 1. Schematic representation of gas phase methanol photooxidation system.
TiO2 Anatase
1wt.% Pd/TiO P90
2
TiO P90
2
1wt.%Pd-Cu(9-1)/TiO P90
2
1wt.%Pd-Cu(3-1)/TiO P90
2
1wt.%Pd-Cu(1-1)/TiO P90
2
TiO2
Rutile
Pd0
PdO
Cu(1)/TiO P90
2
f
e
d
c
b
a
20
25
30
35
40
45
50
55
60
65
70
75
80
2
θ
(deg.)
Fig. 2. XRD patterns of different photocatalysts: (a) 1 wt.% Pd–Cu(1-1)/TiO2, (b) 1 wt.% Pd–Cu(3-1)/TiO2, (c) 1 wt.% Pd–Cu(9-1)/TiO2, (d) TiO2 P90, (e) 1 wt.% Pd/TiO2 P90 and
(f) Cu(1)/TiO2.
3. Results and discussion
bimetallic photocatalysts were studied by XRD. Fig. 2 shows the
diffraction peaks patterns of photocatalysts prepared by sonopho-
todeposition (SPD) method with different atom content Pd/Cu
ratios of 9:1, 3:1 and 1:1. For comparison a pattern of monometal-
lic 1 wt.% Pd/TiO2 P90 photocatalysts and a pattern of commercial
Evonik bare TiO2 P90 were added. The average size of anatase and
rutile crystallites for our photocatalysts, estimated by using the
Scherrer equation, were about 13 and 25 nm, respectively (Table 1).
3.1. Physico-chemical properties of the photocatalysts
In terms of photocatalyst specific surface areas (ca. 104 m2g−1),
pore volume (ca. 0.47 mLg−1), and pore diameter (ca. 15 nm,
mesoporous), all materials showed very similar textural features.
Determination of size and crystalline phase composition of the

P. Lisowski et al. / Journal of Molecular Catalysis A: Chemical 411 (2016) 247–256
251
Fig. 3. DR UV–vis spectra of all tested photocatalysts: (a) TiO2 P90, (b) Cu(1)/TiO2, (c) 1 wt.% Pd–Cu(9-1)/TiO2, (d) 1 wt.% Pd/TiO2 P90, (e) 1 wt.% Pd–Cu(3-1)/TiO2, and (f)
1 wt.% Pd–Cu(1-1)/TiO2. Inset: photocatalyst band gap, Eg, calculations.
Table 1
Structural and optical properties of all tested photocatalysts prepared by the SPD method.
Photocatalyst
DR UV–vis
XRD
Egap (eV)
Absorption threshold (nm)
Crystallite size (nm)
Crystal phase (%)a
Diameter of Pd0 (nm)
Diameter of PdO (nm)
1 wt.%Pd Cu(9-1)/TiO2 P90
1 wt.%Pd-Cu(3-1)/ TiO2 P90
1 wt.%Pd-Cu(1-1)/ TiO2 P90
1 wt.%Pd/TiO2 P90
Cu(1)/TiO2 P90
TiO2 P90
2.53
2.40
2.40
2.51
2.93
2.96
490
516
516
493
423
418
14
13
13
12
13
12
25
27
26
26
25
23
A (55) R (45)
A (57) R (43)
A (59) R (41)
A (74) R (26)
A (78) R (22)
A (84) R (16)
16
24
25
24
–
11
11
11
–
–
–
–
a
A stands for anatase and R for rutile.
Additionally, from the XRD patterns, peaks of palladium oxide
phase (PdO: 2ꢁ = 34◦) and metallic palladium (2ꢁ = 40.1◦) were
detected for all bimetallic photocatalysts prepared by SPD method
[24–26]. The crystallite sizes of palladium oxide in all bimetallic
photocatalysts were approximately 11 nm and the crystallite sizes
of palladium metal were two-fold larger (ca. 22 nm) for the pho-
tocatalysts prepared by SPD method (Table 1). Metallic Pd and
palladium oxide PdO were obtained by SPD method, although, dur-
ing the synthesis, calcination in air flow at 300 ◦C for 4 h was the
last step. It is supposed that the role of sonication is crucial in the
synthesis and for the explanation of our observations. Herein, sono-
chemically generated radicals, like H• from sonolysis of water or
some secondary radical species, are considered to act as reduc-
tants [5–8]. It is believed that these radicals can combine with
oxygen atoms from the palladium (II) acetyloacetonate and there-
fore remove them from the organometallic precursor of palladium.
At the same time, some part of metal ions can be reduced by these
ultrasound-originated radicals and photo-electrons induced in the
photodeposition process. Next, the material is subjected to calci-
nation in air at 300 ◦C with the aim of removing the organic ligand
of the palladium precursor. In the presence of carbon, remaining in
the precursor, the oxygen from the air flow is consumed and pal-
ladium is reduced by the residual carbon. It was observed the XRD
signal of PdO only for bimetallic systems (we do believe that the
presence of copper could retard the total reduction of palladium).
XRD patterns of photocatalysts do not show any clear signs of the
presence of metallic Cu or CuO phases, probably due to the low
amount (below the XRD detection limit) or the high dispersion of
the copper loaded. The XRD spectra of bimetallic catalysts also do
not show any signal that can be attributed to the Pd–Cu alloy phase.
The diffuse reflectance UV–vis spectra (Fig. 3) of synthesized
samples showed the extension of the absorption band to the visible
region (red shift) and a significant enhancement of light absorption
at a wavelength of 418–516 nm. The results obtained indicated that
visible-light absorption of TiO2 P90 prepared by the SPD method
was significantly improved by introducing Pd and Cu nanoparticles
(Eg ≈ 2.4 eV, absorption threshold ꢀ ≈ 516 nm; Table 1).
The absorption band edge is strongly related to the Cu and
Pd nanoparticle size, shape, and Schottky barrier within the
Pd–Cu/TiO2 interfaces in the photocatalyst samples. For sample
1 wt.% Pd/TiO2 P90, sharp absorbance edges occur at a wavelength
of ꢀ ≈ 493 nm, and for Cu(1)/ TiO2 P90 the absorbance edges occur at
a wavelength of ꢀ ≈ 423 nm. The energy values (correlation coeffi-
cient R2 > 0.99) corresponding to the forbidden energy (Eg) for each
material were calculated and are reported in Table 1. The value for
pure TiO2 P90 and Cu(1)/TiO2 were 2.96 eV and 2.93, respectively,

252
P. Lisowski et al. / Journal of Molecular Catalysis A: Chemical 411 (2016) 247–256
Table 2
XPS results for all tested photocatalysts prepared by the SPD method.
Photocatalyst
1 wt.% Pd–Cu(9-
1)/TiO2P90
1 wt.% Pd–Cu(3-
1)/TiO2P90
1 wt.% Pd–Cu(1-
1)/TiO2P90
1 wt.%
Pd/TiO2P90
Cu(1)/TiO2P90
TiO2P90
commercial
Cu 2p BE eV (at%)
Pd 3d BE eV (at%)
Cu0
Pd0
PdO
Ti4+
Ti3+
932.1 (0.05)
335.1 (0.03)
336.2 (0.07)
458.6 (23.60)
–
284.5 (11.28)
285.9 (2.66)
287.4 (1.91)
–
932.2 (0.12)
335.1 (0.04)
336.3 (0.07)
458.8 (18.99)
–
284.5 (15.31)
285.9 (5.84)
287.5 (1.36)
289.1 (0.93)
0.006
932.4 (0.63)
334.9 (0.09)
336.3 (0.07)
458.8 (24.02)
456.8 (3.24)
284.5 (5.47)
285.8 (1.80)
287.4 (0.99)
288.7 (0.94)
0.006
–
932.4 (0.21)
–
–
458.7 (22.1)
456.5 (1.01)
284.5 (9.16)
285.7 (3.93)
287.4 (1.59)
288.8 (0.86)
–
–
–
–
334.9 (0.06)
–
458.7 (23.12)
–
284.5 (5.27)
285.8 (4.00)
287.4 (0.93)
289.1 (1.33)
0.002
458.8 (21.83)
–
Ti 2p BE eV (at%)
C 1s BE eV (at%)
C
C
C
O
C
OH
O
284.5 (6.32)
285.9 (4.29)
287.4 (2.73)
289.8 (1.36)
–
C
O
Pd/Ti atomic ratio
Cu/Ti atomic ratio
0.004
0.002
0.006
0.023
–
0.009
–
whereas the values for the materials containing Cu and Pd on the
surface were between 2.40 and 2.53 eV; this indicates that there
was a slight displacement of the Eg of TiO2 to lower energy val-
ues when Cu and Pd particles were deposited on its surface. It has
been reported by Kamat that contact of metal with the semicon-
ductor indirectly influences the energetics and interfacial charge
transfer processes in a favourable way [27]. Electron accumula-
tion increases the Fermi level of the nanoparticle to more negative
potentials and the resultant Fermi level of the composite shifts
closer to the conduction band (CB) of the semiconductor. There-
fore, the involved edge energy Eg in electron transfer from TiO2 to
the metallic nanoparticles is lower than that of bare TiO2 and the
addition of Pd and Cu leads to enhanced absorption of light in the
visible region by TiO2.
X-ray photoelectron spectroscopy (XPS) analyses were carried
out to determine the chemical and electronic surface structure of
formed nanoparticles and the valence states of the photocatalysts
surface components. The binding energies (BEs) and atomic surface
concentrations of all tested photocatalysts determined by XPS are
given in Table 2. The Cu 2p, Pd 3d and Ti 2p binding energies (BEs)
corresponding to oxidation states are listed in Table 2. In all sam-
ples the dominant peak of Ti 2p is located at 458.8 0.08 eV and
clearly corresponds to Ti4+ in TiO2 structure [28–29]. The binding
energy of Pd 3d in bimetallic photocatalysts were essentially simi-
lar, suggesting that electronic structure of the surface Pd atoms was
not changed in the presence of surrounding neighbours. The pres-
ence of PdO on the surface can be ascribed to the coppers’ retarding
effect in the total palladium reduction process, and also due to the
easy oxidation of Pd upon contact with air at room temperature
as was observed by Herzing et al. [30] that some surface oxide
layer of PdO persisted even after the reduction treatment. It should
be noted that in all bimetallic photocatalysts studied, the bind-
ing energy (BE) signal from Cu0 has shifted to higher (the higher
the amount of copper the larger the shift) values which can be
explained by a charge transfer from Cu0 (presence of Cu␴+) to titania
support (SMSI effect), especially for the best performing photo-
catalyst 1 wt.% Pd–Cu(1-1)/TiO2 P90 where the surface reduction
of Ti4+ → Ti3+ is clearly seen (Table 2). The (Pd/Ti) atomic surface
ratio is comparable in all bimetallic photocatalysts and a slight
increases after the addition of copper is observed. This increases
account a surface coverage by palladium species. On the other hand
significant changes in the (Cu/Ti) was observed in all bimetallic
photocatalysts confirming the surface enrichment in Cu due prob-
ably to the subsurface copper migration onto the titania surface
(2.6-fold higher Cu/Ti atomic ration for 1 wt.% Pd–Cu(1-1)/TiO2 P90
in comparison with Cu(1)/TiO2P90, both with the same nominal
amount of Cu). Additionally, following the comparison between
these two photocatalysts, it was also observed approx. 3-fold higher
concentration of Cu␴+ and Ti3+ surface species for the best photocat-
alytically performing 1 wt.% Pd–Cu(1-1)/TiO2 P90 material (Table 2,
Figs. 6 and 8).
Fig. 4. HRTEM images of photocatalysts (a) 1 wt.% Pd–Cu(3-1)/TiO2 P90, (b) 1 wt.%
Pd–Cu(1-1)/TiO2 P90 and (c) 1 wt.% Pd/TiO2 P90.
Looking for more evidence regarding the nature of the deposited
palladium and copper particles, high-resolution transmission elec-
tron microscopy (HRTEM) measurements were employed. Selected
micrographs of Pd–Cu/TiO2 and Pd/TiO2 samples are shown in
Fig. 4. As it can be observed in Fig. 4a–b in each system palladium
and copper particles mostly form spherically shaped agglomera-
tions with a diameter from dozen to ca. 100 nm. From the average
particle size of TiO2 P90, about 12 nm based on XRD measurements,
the tendency to form aggregates is natural. All dark shadows in the
HRTEM micrographs can be considered as metal-containing parti-
cles. Since both Pd and PdO were present in the sample according to
XRD (copper was not detected probably due to the high dispersion),
the darker areas are most likely attributable to metal presence, as
an effect of denser material. It is interesting to note that all metal-
containing species are always found in close contact especially for
the best performing photocatalyst (Fig. 4b). However, it should

P. Lisowski et al. / Journal of Molecular Catalysis A: Chemical 411 (2016) 247–256
253
Fig. 5. TPO experiment of carbon deposited on 1 wt.% Pd–Cu(1-1)/ TiO2P90 photocatalyst surface after 120 min of photocatalysis.
Fig. 6. Profiles of photocatalytic methanol conversion as a function of light irradiation time over all tested photocatalysts.
Fig. 7. Selectivity to carbon dioxide as a function of light irradiation time over all tested photocatalysts.

254
P. Lisowski et al. / Journal of Molecular Catalysis A: Chemical 411 (2016) 247–256
Fig. 8. Selectivity to methyl formate as a function of light irradiation time over all tested photocatalysts.
be pointed out that no core shell structure was obtained, but the
domains are located next to each other as illustrated in Figure 4a–b.
the photocatalyst surface and then cooled down to room temper-
ature from where a heating ramp rate of 10 ◦C min−1 was used up
to 500 ◦C in 25 mL min−1 of air flow and GC on-line analysis of CO2
(oxidation product) was monitored (Fig. 5). This TPO experiment
showed the presence of carbon deposits on the surface of the pho-
tocatalysts in less than 0.07% of the total amount of methanol after
2 h of photocatalysis.
3.2. Photocatalytic activity and stability
Photocatalytic oxidation of methanol in gas phase was chosen
as the test reaction to evaluate photocatalytic properties of all pre-
pared materials. After 2 h of light irradiation were identified CO2
and methyl formate as the only two reaction products according to
the following reactions (Eq. (1) and (2)):
Fig. 6 displays the photocatalytic conversion of methanol over
monometallic 1 wt.%Pd/TiO2 P90 and bimetallic 1 wt.%Pd–Cu/TiO2
P90 photocatalysts as a function of light irradiation time, together
with TiO2 P90 (Evonik) photocatalyst for comparison. It is clearly
seen that modification of TiO2 with palladium and copper apply-
ing sonophotodeposition method resulted in a great enhancement
of photocatalytic activity and selectivity to methyl formate. Titania
itself was active (ca. 33% methanol conversion, Fig. 6) only towards
total oxidation of methanol to carbon dioxide under UV irradiation.
It should be noted that the 1 wt.%Pd/TiO2 P90 exhibited the highest
methanol conversion among the all tested photocatalysts but with-
out methyl formate production and with the highest grade of total
mineralization (83%, Fig. 7) after 2 h of light irradiation. Among all
tested materials the best selectivity (approx. 80%, Fig. 8) to methyl
formate MF with good methanol conversion of 53% (this conver-
sion was markedly increased at the initial stage of the reaction (first
15 min) and then slowly decreased with time (approx. 8% decrease)
after 2 h of illumination, Fig. 6) was obtained for 1 wt.%Pd–Cu(1-
1)/TiO2P90 prepared by sonophotodeposition method as a function
of light irradiation. It was also observed that photocatalysts con-
taining Cu–Pd nanoparticles supported on TiO2 P90 exhibit much
lower mineralization rate to carbon dioxide under applied reac-
tion conditions (Fig. 7). Furthermore, we obtained worse results for
the 1 wt.% Pd–Cu(1-1)/TiO2P90/PD (results not shown here) photo-
catalyst prepared by only the photodeposition method (methanol
conversion of 28%, slowly decreased with time with approx. 5%
deactivation after 2 h of illumination and selectivity to MF and CO2
of 64% and 36%, respectively).
2CH3OH + 3O2 → 2CO2 + 4H20
2CH3OH + O2 → HCOOCH3 + H20
(1)
(2)
Initially, three control experiments were applied: (1) photol-
ysis upon UV illumination in the presence of methanol in the
air flow and without photocatalyst, (2) photocatalytic methanol
oxidation in the presence of the 1 wt.% Pd–Cu(1-1)/TiO2 pho-
tocatalyst without oxygen (argon instead of air), and (3) the
thermal effect (up to 100 ◦C) in the dark, in the presence of
the photocatalyst and methanol in the air flow. Additionally, for
the best performing photocatalyst (1 wt.% Pd–Cu(1-1)/TiO2 P90),
temperature-programmed oxidation (TPO) of coke deposited on its
surface after 2 h of photocatalysis was conducted (Fig. 5).
In the photolysis experiment (absence of photocatalyst), very
low conversion of methanol (<3%) was observed, thus confirm-
ing that the reaction is really enabled by a photocatalytic process.
Moreover, the catalyst was not active under thermal conditions (up
to 100 ◦C) in the dark. Therefore, the conclusion is that this reac-
tion depends on the presence of both light and photocatalyst. In the
absence of oxygen (air), 1 wt.%Pd–Cu(1-1)/TiO2 P90 photocatalyst
exhibited low methanol conversion (<10%) with slight deactivation
and the highest selectivity to methyl formate (80%, Figs. 6 and 8).
We believe that in the absence of oxygen (air), selective oxidation
takes place through TiO2 (in 1 wt.%Pd–Cu(1-1)/TiO2 P90) lattice
oxygen atoms and that oxygen (supplied in air flow) was needed
only to replenish the produced oxygen vacancies on the TiO2 lat-
tice. When gas-phase oxygen was present, the oxidation rate of
methanol was greatly improved (Fig. 6) suggesting the important
role of adsorbed oxygen.
Since the activity tests described above were carried out with
high purity gases, the inhibition by impurities in the gases is not the
primary source for the deactivation. Nevertheless, numerous stud-
ies indicate that a certain degree of humidity, usually exceeding
the amount produced by the oxidation of the organics, is neces-
sary to maintain hydroxylation and to avoid the blockage of the
TiO2 surface by partially oxidized products [31–32]. Moreover, the
hydrophilicity of the TiO2 surface will influence the adsorption and
desorption behavior of the reactants and products. Therefore the
use of hydrocarbons with different molecular functionalities can
help to get more insight into the charge transfer mechanisms. The
The TPO experiment was used to measure the amount of
“organic residues” potentially remaining on the surface of the best
performing photocatalyst (1 wt.% Pd–Cu(1-1)/TiO2 P90) after the
reaction. After 2 h of light irradiation, 1 wt.%Pd–Cu(1-1)/TiO2P90
photocatalyst was heated first at 100 ◦C for 2 h in 25 mL min−1 of
helium flow to remove all physisorbed reagents and products from

P. Lisowski et al. / Journal of Molecular Catalysis A: Chemical 411 (2016) 247–256
255
active site for a given product could be a certain atomic arrange-
ment of metal atoms, the surface atom in a certain oxidation
state, bimetallic surface site with a specific composition, or the
interface between the metal nanoparticle and the oxide support.
Under reaction conditions, active sites may reconstruct due to the
light effect, or be blocked by strongly adsorbed intermediates. Our
results potentially imply that the photocatalysts are deactivated
by accumulation of surface species formed during the reaction
[33]. Methanol for example will adsorb strongly to the surface and
will form weaker adsorbed intermediates [34]. Decomposition of
methanol proceeds via formation of methoxy (CH3O) species, fol-
lowed by either C H or C O bond scission. C H bond scission
leads to formaldehyde (CH2O), formyl (CHO) and finally CO, fol-
lowed by oxidation to CO2 in the presence of adsorbed oxygen. O
and C atomic co-adsorbates may play a critical role under reaction
conditions. In the presence of oxygen, surface and bulk oxidation
may occur, whereas carbon can accumulate on the surface, but also
in the subsurface and bulk region. We believe (more research is
needed to prove this) that these carbon species adsorb on [Pd—
selective oxidation of methanol. In such specific environment, the
␴+
[Pd–Cu ↔ Ti3+Ox] strong interactions are quite probable, espe-
cially after optimization of one of the most important parameters
such as the appropriate metals’ composition as it was observed for
the best performing photocatalyst 1 wt.%Pd–Cu(1-1)/TiO2 P90, but
still this hypothesis needs more research to be totally proven.
4. Conclusions
Surface modification of TiO2 with copper and palladium has
been successfully carried out using sonophotodeposition method
with TiO2 P90 as support. Compared with the commercial refer-
ence, bare TiO2 P90, UV–vis diffuse reflectance spectra indicate that
the photoabsorption of Pd–Cu/TiO2 photocatalysts is extended to
the visible region. Bimetallic systems (especially 1 wt.%Pd–Cu(1-1)
on TiO2 P90) were nicely selective (approx. 80% of methyl formate)
with good activity (53% conversion) and showing slow deactivation
with time to approx. 8% conversion decrease after 2 h of illumina-
tion. As a matter of fact, the majority of the benefit comes from
advantageous physicochemical properties of the Pd–Cu/TiO2 P90
photocatalysts, in terms of size of its metal particles, their loca-
tion relative to the support and their mobility, owing to an optimal
extent of catalysts’ strong metal-support interaction effect. This
photocatalyst may also be slightly deactivated by accumulation of
surface carbonaceous species formed during the reaction.
Sonophotodeposition method is a green methodology that can
be used to prepare well-defined bimetallic surfaces on semicon-
ductor supports with great promise for catalytic applications, in
which selectivity can be tuned through adjustment of the surface
composition.
␴+
Cu ↔ Ti3+Ox] active-selective site systematically poisoning the
SMSI effect observed for our best photocatalyst. Under such con-
ditions the contribution of C O bond scission becomes stronger,
leading to the formation of carbonaceous overlayers and to catalyst
deactivation [35].
Nimlos et al. [36] reported that adsorbed maximum amount
on TiO2 and the adsorption equilibrium constant followed the
sequence: acetic acid > ethanol > acetaldehyde. It indicated that
carboxylic acid was strongly adsorbed onto the TiO2 surface. Con-
sequently, the deactivating species could be the carboxylic acid.
The same decay of photocatalytic activity was observed when 1-
propanol and propionaldehyde were photo-oxidized; the species
responsible for the deactivation may be the same in both cases.
It can be assumed that the deactivating species may be propionic
acid, acetic acid or formic acid. The amount of intermediates was
higher at lower mineralization extent, therefore more partial oxi-
dized species were adsorbed on the catalyst surface to cause the
loss of photocatalytic activity. Additionally, selectivity of hetero-
geneous photocatalytic reactions is ultimately determined by the
relative concentrations of active sites for different reaction path-
ways. The deactivation may also to some extent be due to water
inhibition generated in the mineralization reaction (production of
CO2 and H2O).
Acknowledgements
This work was partially supported by the National Sci-
ence Centre (NCN) in Poland within research project DEC-
2011/01/B/ST5/03888. We would also like to thank the Institute
of Physical Chemistry of PAS and the COST Association (Action
FP1306) for supporting the dissemination of this work.
References
[1] Y. Qu, X. Duan, Chem. Soc. Rev. 42 (2013) 2568–2580.
[2] M.D. Hernandez-Alonso, F. Fresno, S. Suareza, J.M. Coronado, Energy. Environ.
Sci. 2 (2009) 1231–1257.
[3] A. Fujishima, X. Zhang, D.A. Tryk, Surf. Sci. Rep. 63 (2008) 515–582.
[4] J.C. Colmenares, R. Luque, Chem. Soc. Rev. 43 (2014) 765.
[5] H. Xu, B.W. Zeiger, K.S. Suslick, Chem. Soc. Rev. 42 (2013) 2555.
[6] J.C. Colmenares, J. Nanosci. Nanotechnol. 13 (2013) 4787.
[7] J.H. Bang, K.S. Suslick, Adv. Mater. 22 (2010) 1039.
The addition of copper improve the selectivity to methyl formate
(>70%) in the monometallic Pd/TiO2 system with slowly decreasing
methanol conversion. In contrast, 1%Pd/TiO2 P90 and TiO2 P90 are
non-active in the selective methanol oxidation to methyl formate
but are active and selective to CO2 formation (total mineraliza-
tion). Such results suggest the advantages of the bimetallic Cu–Pd
catalysts over the monometallic in the selective methanol conver-
sion. The role of noble metals on semiconductor surfaces seems to
be very well known. Many researchers claim that they act as effi-
cient catalysts for electron transfer reactions and also decrease the
recombination probability of photoholes with their counterparts
(electrons) and therefore increase the fraction of photoholes avail-
able for oxidizing interfacial charge-transfer reactions [37]. In air
[8] K.S. Suslick, Science 247 (1990) 1439–1445.
[9] M. Lewandowski, D.F. Ollis, Appl. Catal. B Environ. 43 (2003) 309–327.
[10] M.L. Sauer, D.F. Ollis, J. Catal. 163 (1996) 215.
[11] J. Peral, D.F. Ollis, J. Catal. 136 (1992) 554.
[12] S.A. Larson, J.L. Falconer, Appl. Catal. B Environ. 4 (1994) 325.
[13] A.V. Vorontsov, E.N. Kurkin, E.N. Savinov, J. Catal. 186 (1999) 318–324.
[14] R. Méndez-Román, N. Cardona-Martinez, Catal. Today 40 (1998) 353.
[15] A. Huang, L. Cao, C. Jie, S. Franz-Josef, L.S. Steven, T.N. Obee, S.O. Hay, J.D.
Freihaut, J. Catal. 188 (1999) 40.
[16] J.C. Colmenares, P. Lisowski, D. Łomot, O. Chernyayeva, D. Lisovytskiy,
ChemSusChem 8 (2015) 1676–1685.
␴+
atmosphere, the electron trapped by [Pd—Cu ↔ Ti3+Ox] active-
[17] J.C. Colmenares, P. Lisowski, Polish Patent Appl. (2013) P-405094.
[18] S. Brunauer, P.H. Emmett, E. Teller, J. Am. Chem. Soc. 60 (1938) 309–319.
[19] E.P. Barrett, L.G. Joyner, P.P. Halenda, J. Am. Chem. Soc. 73 (1951) 373–380.
[20] B.D. Cullity, S.R. Stock, Elements of X-ray Diffraction, 3rd ed., Prentice Hall
Inc., Upper Saddle River, NJ, 2001.
[21] R.A. Spurr, H. Myers, Anal. Chem. 29 (1957) 760–762.
[22] S. Sakthivel, H. Kisch, Angew. Chem. Int. Ed. 42 (2003) 4908–4911.
[23] H.-S. Lee, C.-S. Woo, B.-K. Youn, S.-Y. Kim, S.-T. Oh, Y.-E. Sung, H.-I. Lee, Top.
Catal. 35 (2005) 3–4.
[24] K. Mandal, D. Bhattacharjee, P.S. Roy, S.K. Bhattacharya, S. Dasgupta, Appl.
Catal. A 492 (2015) 100–106.
[25] M.-S. Kim, S.-H. Chung, C.-J. Yoo, M.S. Lee, I.l.-H. Cho, D.-W. Lee, K.-Y. Lee,
Appl. Catal. B Environ. 142–143 (2013) 354–361.
selective sites is transferred to O2 and the superoxide anion is
•
␴+
formed, O2 −. After releasing the electron [Pd—Cu ↔ Ti3+Ox] can
trap another one and the whole process repeats leaving the photo-
hole free and capable to oxidize selectively methanol. However,
we should remember that the catalytic behavior of a bimetallic
material may depend on both its electronic and geometric struc-
ture [38]. As one can see, the high selectivity to methyl formate
depends highly on metals’ loading, oxidation state of the metals
and their interfacial behavior on titania surface. As was observed
by XPS measurements, SMSI effect can have an important role in the

256
P. Lisowski et al. / Journal of Molecular Catalysis A: Chemical 411 (2016) 247–256
[26] O.S.G.P. Soares, M.F.R. Pereira, J.J.M. Órfão, J.L. Faria, C.G. Silva, Chem. Eng. J.
251 (2014) 123–130.
[27] P.V. Kamat, J. Phys. Chem. B 106 (2002) 7729–7744.
[28] C. Xua, Y. Liua, J. Wanga, H. Genga, H. Qiu, J. Power Sources 199 (2012)
124–131.
[29] J.F. Moulder, W.F. Stickle, P.E. Sobol, K.D. Bomben, Hand-book of X-ray
Photoelectron Spectroscopy, in: J. Chastain (Ed.), PerkinElmer Corporation
Physical Electronics Division, USA, 1992.
[30] A. Herzing, A.F. Carley, J.K. Edwards, G.J. Hutchings, C.J. Kiely, Chem. Mater.
(2015) 20.
[31] M.M. Ameen, G.B. Raupp, J. Catal. 184 (1999) 112–122.
[32] A.J. Maira, K.L. Yeung, J. Soria, J.M. Coronado, C. Belver, C.Y. Lee, V. Augugliaro,
Appl. Catal. B Environ. 29 (2001) 327–336.
[33] T.N. Obee, R.T. Brown, Environ. Sci. Technol. 29 (1995) 1223–1231.
[34] A.B. Anderson, H.A. Asiri, Phys. Chem. Chem. Phys. 16 (2014) 10587.
[35] M. Baumer, J. Libuda, K.M. Neyman, N. Rosch, G. Rupprechterz, H.-J. Freund,
Phys. Chem. Chem. Phys. 9 (2007) 3541–3558.
[36] M.R. Nimlos, E.J. Wolfrum, M.L. Brewer, G. Bintner, Environ. Sci. Technol. 30
(1996) 3102–3110.
[37] A.A. Ismail, S.A. Al-Sayari, D.W. Bahnemann, Catal. Today 209 (2013) 2–7.
[38] B. Coq, F. Figueras, J. Mol. Catal. A: Chem. 173 (2001) 117–134.