Efficient approach for simultaneous CO and H2 production via photoreduction of CO2 with water over copper nanoparticles loaded TiO2

Article abstract of DOI:10.1016/j.apcata.2016.05.027

Photoreduction of CO₂ with water to alternative fuels using low cost meterials is an efficient approach to mitigate CO₂. The present work describes a series of copper nanoparticles (Cu-NPs) loaded on TiO₂, here copper nanoparticles were separately prepared by chemical reduction method, and then loaded on freshly synthesized TiO₂. Oxidized and reduced forms of Cu/TiO₂ nanoparticles were obtained by calcination under air and H₂ respectively. Prepared catalysts were characterized by XRD, SEM, EDX, HR-TEM, XPS, BET surface area analysis, TPR, UV-vis diffuse reflectance spectroscopy and spectrofluorometry. Catalytic potential of prepared catalysts was checked for photocatalytic reduction of CO₂ with water vapor under UV-C irradiation. The effect of oxidation state and loading amount of Cu as well as irradiation wavelength was also investigated. CO and H₂ were observed as photocatalytic products. Among all the photocatalysts, 1% Cu-Red/TiO₂ exhibited the highest photoactivity, with an initial rate of CO and H₂ production of 334 and 452 μmol g^-1 h^-1 respectively, under UV-C light. The photocatalytic role of the Cu-NPs surface giving contribution to overall reaction was confirmed by using Cu-NPs, 1% Cu-Oxi/SiO₂ and 1% Cu-Red/SiO₂. A plausible mechanism of CO₂ reduction is suggested and computationally confirmed. Reaction mechanism includes stages of CO₂ adsorption at the Cu-TiO₂ interface, reduction into HCO₂, and dissociation into adsorbed hydroxyls and CO. The most energy demanding step was found to be CO desorption that could be facilitated by 254 nm quanta.

Full text of DOI:10.1016/j.apcata.2016.05.027

Accepted Manuscript
Title: Efficient approach for simultaneous CO and H2
production via photoreduction of CO with water over copper
2
nanoparticles loaded TiO2
Author: Nikita Singhal Asgar Ali Alexander Vorontsov
Chandrashekar Pendem Umesh Kumar
PII:
S0926-860X(16)30305-2
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.apcata.2016.05.027
APCATA 15891
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
31-12-2015
23-5-2016
29-5-2016
Please cite this article as: Nikita Singhal, Asgar Ali, Alexander Vorontsov,
Chandrashekar Pendem, Umesh Kumar, Efficient approach for simultaneous CO and
H2 production via photoreduction of CO2 with water over copper nanoparticles loaded
TiO2, Applied Catalysis A, General http://dx.doi.org/10.1016/j.apcata.2016.05.027
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.

Efficient approach for simultaneous CO and H2 production via
photoreduction of CO with water over copper nanoparticles loaded TiO2
2
[
ab]
[a]
[c]
[d]
Nikita Singhal , Asgar Ali , Alexander Vorontsov Chandrashekar Pendem and Umesh
[
ab]
Kumar
*
umeshk@iip.res.in
[
[
[
[
a] Chemical Science Division, CSIR•Indian Institute of Petroleum, Dehradun•248005, India
b] AcSIR-Academy of Scientific & Innovative Research New Delhi, India
c] Boreskov Institute of Catalysis, Novosibirsk 630090, Russia
d] Chandrashekar Pendem, Refinery Technology Division,
CSIR•Indian Institute of Petroleum, Dehradun•248005, India
Corresponding Author
*
1

Graphical Abstract
2

Highlights



Cu nanoparticles (NPs) were synthesized via chemical reduction route.
Cu-NPs were loaded on lab synthesized TiO NPs to form Cu/TiO .
2
2
The Cu-NPs loading amount and oxidation state controls the copper particle size and the
products yield.


CO and H were produced under UV irradiation from CO and water vapor over Cu/TiO .
2
2
2
-1 -1
The 1% Cu-Red/TiO photocatalyst showed maximum 334 µmolg h CO production
2
during CO photoreduction.
2


CO desorption was theoretically demonstrated to be the rate limiting step.
The catalyst structure and its effect on the CO reduction were elucidated via quantum
2
chemical modeling.
3

ABSTRACT
Photoreduction of CO with water to alternative fuels using low cost metals is an efficient
2
approach to mitigate CO . The present work describes a series of copper nanoparticles (Cu-NPs)
2
loaded on TiO , here copper nanoparticles were separately prepared by chemical reduction
2
method, and then loaded on freshly synthesized TiO . Oxidized and reduced forms of Cu/TiO
2
2
nanoparticles were obtained by calcination under air and H respectively. Prepared catalysts were
2
characterized by XRD, SEM, EDX, HR-TEM, XPS, BET surface area analysis, TPR, UV-vis
diffuse reflectance spectroscopy and spectrofuorometry. Catalytic potential of prepared catalysts
was checked for photocatalytic reduction of CO with water vapour under UV-C irradiation. The
2
effect of oxidation state and loading Cu as well as irradiation wavelength was also investigated.
CO and H were observed as photocatalytic products. Among all the photocatalysts, 1% Cu-
2
Red/TiO exhibited the highest photoactivity, with an initial rate of CO and H production of 334
2
2
-1 -1
and 452 µmolg h respectively, under UV-C light. The photocatalytic role of the Cu-NPs
surface giving contribution to overall reaction was confirmed by using Cu-NPs, 1% Cu-Oxi/SiO2
and 1% Cu-Red/SiO . A plausible mechanism of CO2 reduction is suggested and
2
computationally confirmed. Reaction mechanism includes stages of CO adsorption at the Cu-
2
TiO interface, reduction into HCO , and dissociation into adsorbed hydroxyls and CO. The most
2
2
energy demanding step was found to be CO desorption that could be facilitated by 254 nm
quanta.
Keywords: Copper nanoparticles, Cu/TiO , photocatalyst, CO photoreduction, CO and H
2
2
2
production, DFT, PM6, clusters.
4

1
. Introduction
The continuous increment of anthropogenic CO in environment is responsible for the climatic
2
change and global warming i.e. an increase in global temperature. Moreover it is an opportunity
filled with challenges to employ anthropogenic CO as feedstock for chemicals and fuels
2
synthesis [1]. CO can be used for production of CO, CH , HCHO, HCOOH and CH OH by
2
4
3
reduction [2]. CO reduction can be done in many ways as electrochemical reduction [3],
2
hydrogenation at high temperature [4], light induced photochemical reduction [5]. The
photocatalytic reduction of CO is the best tool, because it uses abundant sources of energy and
2
water to reduce CO to hydrocarbons [6]. Several photocatalysts have been reported for CO
2
2
reduction under UV irradiation like ZrO , ZnO, TiO , ZnGe O , Ta O , SnO and many more
2
2
2
4
2
5
2
[
7]. Development of a safe, abundant and cost effective catalyst is a considerable significance in
solving current energy issues [8-10]. The TiO is well explored photocatalyst due to its cost
2
effectiveness, abundance, nontoxicity, stability and suitable electronic and optical properties [11-
1
2]. The anatase phase of TiO (3.2 eV band gap) is most active under UV radiation but owing to
2
massive recombination of generated electron hole pair, some modifications are required to entrap
hole or electron to carry out desired reaction [13]. The photogenerated electron trap by
impregnated metal at the surface of semiconductor can delay recombination process and thus
provide a direct photoreduction of CO [14]. Due to the low cost and availability, copper would
2
be a better choice to modify TiO . Many researchers had reported copper modified TiO catalysts
2
2
for enhanced photocatalytic activities and also photocatalytic reduction of CO [15-19]. Liu et al.
2
have synthesized copper-loaded titania photocatalysts in one pot reaction and observed
enhancement in CO reduction to CO than with pure TiO [20]. Chen et al recently reported CO
2
2
2
5

-1
hydrogenation to maximum 28.72 μmol g methane over Cu modified TiO under UV-C
2
+
+3
+4
irradiation. They showed presence of Cu and Cu species and ratio of Ti with respect to Ti
enhances the photoactivity [21]. Tan et al. have prepared copper nanoparticles decorated TiO2
-1 -1
nanorod films using electrochemical method for reduction of CO to methane (2.91 ppm g h
2
catalyst) [22]. J. Zhao et al prepared metallic Cu modified black TiO for CO photoreduction
2
2
with water vapor under visible light irradiation. 4% Cu@TiO produces maximum 0.97 μmoles
2
of CO and 0.16 μmoles of methane [23]. The impact of copper dispersion was also studied by
-1 -1
Liu et al. and they have obtained the maximum 24 nmol g h of methane [24].
Herein, we describe the successful synthesis of copper/TiO NPs in which ex-situ synthesized
2
Cu-NPs were loaded on the surface of TiO NPs to get copper/TiO NPs. Copper/TiO NPs
2
2
2
synthesis, characterization, and their use for the CO photoreduction with water vapour to yield
2
CO as major product under UV radiation were studied. The quantitative and qualitative
determination of the CO and H formation was performed by refinery gas analyzer (RGA). The
2
Quantum chemical modelling of catalyst structure and CO reduction was used to derive the
2
reaction mechanism of CO reduction.
2
2
. Experimental
2
.1
Materials
Titanium butoxide (≥97%, Sigma Aldrich), nitric acid (69%, Merck), glacial acetic acid
≥99.8%, Merck UK), absolute ethanol (AR grade, Fischer Chemicals UK), CuCl (AR grade,
(
2
BDH), L-ascorbic acid (AR grade, ≥99.7%, Sigma Aldrich) and HPLC grade water were used as
chemicals.
6

2
.2
Synthesis
a.
Cu-NPs Synthesis
Cu-NPs were prepared by chemical reduction method using ascorbic acid as a mild reducing and
capping agent [25]. In a typical synthesis, 0.2 M aqueous solution of copper chloride was
reduced by adding equal volume of 0.5 M L-ascorbic acid with subsequent stirring for 24 h at
8
0 °C during reaction, the color of reaction mixture changed from green to dark brown. The
resultant reaction mixture was centrifuged at 6000 rpm for 15 min and supernatant was stored for
further use. The copper concentration of Cu-NPs dispersion solution was measured by ICP-AES
and observed as 0.081 M.
b.
TiO Preparation
2
TiO was prepared by a sol-gel method. Here 10.0 g titanium butoxide was dissolved in 40 mL
2
ethanol, then pH of solution was adjusted to < 2 by slow addition of dilute ethanolic nitric acid.
2
0 mL of acetic acid was added to the above mixture to facilitate the formation of homogeneous
colloidal gel, with the help of slowly released water on esterification. The reaction mixture was
stirred in controlled atmosphere for 24 h at room temperature to get homogeneous gel. The
prepared gel was dried in an air oven at 80 °C and grinded well followed by calcination under air
atmosphere at 600 °C for 4 h with a ramp rate of 2 °C/min. After cooling to room temperature,
the prepared white TiO powder was crushed in a mortar to get homogeneous powder.
2
c.
Cu-Oxi/TiO2
The oxidized copper-containing photocatalyst with loading of 1 and 2 wt% on TiO was prepared
2
by mixing 2 g of TiO with 3.92 mL and 7.84 mL of freshly prepared Cu-NPs solution
2
respectively. The above mixture was dried at 80 °C, grinded well and calcined at 600 °C in air
7

for 4 h to get 1% Cu-Oxi/TiO and 2% Cu-Oxi/TiO catalysts. After cooling to room
2
2
temperature, the prepared photocatalysts were crushed in a mortar to get homogeneous powders.
d. Cu-Red/TiO2
The reduced photocatalyst containing 1 and 2 wt% copper on TiO was prepared by following
2
the same loading method as for the Cu-Oxi/TiO followed by reduction under 20% hydrogen at
2
3
50 °C for 3 h. After cooling to room temperature, the prepared 1% Cu-Red/TiO and 2% Cu-
2
Red/TiO catalysts were crushed in a mortar to get homogeneous powders and stored under inert
2
atmosphere.
2
.3
Catalyst Characterization
Catalysts were characterized with different spectroscopic and analytical techniques. X-ray
diffraction (XRD) patterns were obtained on a Bruker D8 Advance X-ray diffractometer using
monochromatic Cu Kα (λ= 1.5418 Å) radiation in 2θ range from 20° to 80° with a step size
0
.02°. Morphological information and elemental mapping of the catalyst particles were obtained
by FEI Quanta 200F scanning electron microscope (SEM) system equipped with energy-
dispersive X-ray (EDX) spectroscopy. Transmission electron microscopic analysis was
performed on FEI model no Technai G2 20 S-TWIN instrument. The HR-TEM analysis was
done on JEM-2010 instrument (JEOL, Japan). The accelerating voltage was 200 kV, resolution
was 1.4 Å. The catalyst particles were loaded on the carbon coated copper grid via dispersion in
ethyl alcohol using ultrasonic processor. The local elemental composition in HR-TEM was
measured by energy-dispersive spectrometer (EDX) produced by Phoenix using Si (Li) detector
with energetic resolution of 130 eV. The solid state UV-Vis diffuse reflectance spectra were
recorded with a Schimadzu 2600 UV/Vis-NIR spectrophotometer using BaSO as the totally
4
8

reflecting standard. X-ray photoelectron spectroscopy (XPS) analysis of materials was carried
out on model S/N-10001 (Prevac Poland) with a VG Scienta-R3000 hemispherical energy
analyzer. H -TPR measurements were carried out using Micromeritics AutoChem II 2920
2
analyzer from Micromeritics, USA to find out the consumption of H by copper oxide loaded at
2
the surface of TiO using TCD detector. Metal dispersion measurements of catalysts were carried
2
out by H -TPR followed by N O passivation and H -TPR. The 0.05 g of Cu loaded TiO catalyst
2
2
2
2
was preheated with 30 mL/min of helium flow at 350 °C for 3h and TPR analysis was carried out
from 50 – 550 °C using 10% H /Ar gas mixture. Then reduced catalyst passivated with 10%
2
N O/He at 90 °C for 20 min and physorbed N O was removed by He gas stream at 50 mL/min
2
2
for 60 min and oxidized surface Cu catalyst reduced by H -TPR at similar condition. The Cu
2
dispersion was calculated as per literature known method [26-27]. The amount of copper present
in freshly prepared Cu-NPs solution was estimated by inductively coupled plasma atomic
emission spectrometer (ICP-AES) model PS 3000 uv, (DRE), Leeman Labs, Inc. (USA)]. The
Cu-NPs solution was mixed with HNO for digestion of copper followed by make up with
3
deionized HPLC water.
2
.4
Photocatalytic Activity Measurements
The photocatalytic reduction of carbon dioxide in the presence of water vapor was carried
out in the reaction assembly as shown in SI-Fig.1. The reaction setup consists of a cylindrical gas
tight reactor (H = 10.5 cm, ID = 5.5 cm) equipped with a gas inlet, outlet, gas sampling port, and
a quartz tube at the centre of reactor for insertion of irradiation source. The flow of nitrogen and
CO from gas cylinders was maintained by mass flow controllers and passed through a saturator
2
filled with water kept at 25 °C. A thin layer of 0.25 g catalyst was spread over the bottom of
reactor. Photoreactor was preconditioned by repeated evacuation and purging with nitrogen
9

(
99.995%, Sigma gas) and final evacuation followed by purging with water saturated CO2
99.999%, Sigma gas) for 1 h.
(
The photocatalyst was illuminated by 8 W UV-C lamp (PEN-RAY, 254 nm, 2.13", Cole-
2
Parmer, 4400 µW/cm ) or 8 W UV-A lamp (PEN-RAY lamp, 350 nm, 2.13", Cole-Parmer, 120
2
2
µW/cm ) or 8 W white visible light cold LED lamp (Siska LED 20 µW/cm ) to drive the
reaction. The reaction progress was monitored through continuous sampling at regular time
intervals and analysis with Refinery Gas Analyzer (RGA) (Agilent 7890) using nitrogen as a
carrier gas. RGA was equipped with a HayeSep Q capillary column, two thermal conductivity
detectors (TCD), and one flame ionization detector (FID). Since this study is focused at CO2
reduction at the gas–solid interface, possible products characteristic of liquid phase reduction
such as methanol, formaldehyde and formic acid were not measured [28-31].
2
.5
Quantum Modeling of Cu/TiO Photocatalyst and CO Reduction
2 2
TiO anatase NP is represented by a small cluster model exposing (101) facet which is
2
known as the most stable anatase surface. Dangling bonds of the terminal oxygen atoms of the
TiO cluster are saturated with hydrogen atoms. Cu-NP was approximated by the cluster
2
containing four atoms. Enthalpy of formation of the photocatalyst, reagents, products and surface
bound intermediates was calculated with PM6 method (MOPAC 2012 software). This method
has been demonstrated to provide reliable results for TiO anatase nanoparticles. Further
2
1
0

description of the models used is given in the section 3.3, devoted to the modeling results and
discussion.
1
1

3
3
. Results and discussion
.1 Catalyst Characterization
X-ray diffraction patterns of TiO prepared by sol-gel route, 2% Cu-Oxi/TiO and 2% Cu-
2
2
Red/TiO are shown in Fig. 1. XRD confirms that prepared catalysts contain only anatase phase
2
of TiO . Major peaks at 25.3°, 37.8°, 48.1° (2θ) correspond to 101, 004, and 200 hkl planes of
2
TiO anatase (JCPDS 21-1272). The characteristic peaks for hkl 002 and 111 plane of CuO are
2
usually observed the region of ca. 36° to 39° (2θ) (JCPDS card no. 65-2309). Here, no
characteristic peaks of copper oxide appeared in XRD pattern of 2% Cu-Oxi/TiO due to the
2
presence of a broad peak of TiO in the range 2θ = 35 - 40°. There is a small peak present at ca.
2
4
3.5° corresponding to 111 hkl plane of copper metal in 2% Cu-Red/TiO (JCPDS 85-1326)
2
[
32].
The surface area of a catalyst is an important factor for better work efficiency. Surface area of all
photocatalysts was determined by low temperature N adsorption following BET surface
2
2
analysis. The specific surface area of TiO is 41.5 m /g. But after loading of Cu-NPs, specific
2
2
2
surface area has slightly decreased 1% Cu-Oxi/TiO (40.2 m /g), 2% Cu-Oxi/TiO (40.1 m /g),
2
2
2
2
1
% Cu-Red/TiO (40.3 m /g) and 2% Cu-Red/TiO (39.7 m /g). The presence of oxi or reduced
2 2
copper does not make significant morphological change on the surface of TiO hence no
2
significant change in surface area.
1
2

Fig. 1. XRD patterns of TiO , 2% Cu-Oxi/TiO and 2% Cu-Red/TiO . Inset reveals XRD peak
2
2
2
having 111 plane of metallic copper.
Electronic diffuse reflectance spectra of TiO and Cu-NPs loaded TiO photocatalysts are shown
2
2
in Fig. 2a. The prepared TiO absorbs light in the range of 380 – 400 nm corresponding to band-
2
1
3

gap of ca. 3.2 – 3.1 eV. There was not much difference between the spectra of TiO and oxidized
2
Cu/TiO indicating that low concentration of CuO did not introduce significant absorption. In
2
2
+
addition one week intensity band was observed in the region of 600 – 800 nm due to Cu in
oxidized Cu/TiO [33]. Reduced Cu/TiO photocatalysts are of dark gray color and show
2
2
considerable absorption in the visible light range. The electronic spectra of freshly prepared Cu-
NPs are shown in Fig. 2b. The surface plasmon resonance of copper is usually observed at 570
nm but for <4 nm particle size this peak become broadened and featureless absorbance [34-35].
The absorption onset of a wide peak occurs at wavelength of about 570 nm and the absorption
flattens at 530 nm. This wide peak corresponds to surface plasmon resonance of variable size
Cu-NPs [36]. The TEM image in the inset of Fig. 2b clearly shows that Cu-NPs are spherical in
shape and in the range of 2.5-10 nm.
1
4

Fig. 2. Electronic spectra of a) TiO and Cu-NPs loaded TiO ; b) Cu-NPs prepared by ascorbic
2
2
acid reduction. The inset reveals TEM of Cu-NPs.
SEM images of synthesized TiO and Cu/TiO catalysts as well as mapping of Cu are shown in
2
2
Fig. 3. SEM image of prepared TiO is shown in Fig. 3a. As expected elemental mapping in Fig.
2
3
a’ does not have the sharp dots for copper while measuring the copper in neat TiO which
2
reveal absence of Cu. Fig. 3b, 3b’, 3c and 3c’ are SEM images of 2% Cu-Oxi/TiO and 2% Cu-
2
Red/TiO and their elemental mapping for copper. Fig. 3b’ and 3c’, reveal the homogeneous
2
dispersion of Cu-NPs on TiO surface. Fig 3a”, 3b” and 3c”, reveal the EDX of TiO , 2% Cu-
2
2
Oxi/TiO and 2% Cu-Red/TiO . The fig 3a” does not have copper peaks while 3b” and 3c” has
2
2
the presence of copper.
1
5

Fig. 3. SEM of a) TiO ; b) 2% Cu-Oxi/TiO ; c) 2% Cu-Red/TiO . Elemental mapping of a’)
2
2
2
back ground noise mapping for copper in TiO ; b’) Cu in 2% Cu-Oxi/TiO ; c’) Cu in 2% Cu-
2
2
Red/TiO . EDX of a”) TiO ; b”) 2% Cu-Oxi/TiO ; c”) 2% Cu-Red/TiO .
2
2
2
2
TEM images of TiO , 2% Cu-Oxi/TiO , 2% Cu-Red/TiO are shown in Fig. 4. The feed stock
2
2
2
Cu-NPs are close to spherical in shape and their size lies in the range of 2.5 - 10 nm (Fig. 2b
inset), the majority of particles are of around 5 nm. TEM results in Fig. 4a show that the
1
6

photocatalyst TiO has the particle size of 15 – 50 nm and HR-TEM image in Fig. 4b
2
demonstrates clear lattice fringes indicating crystalline nature of catalyst. In Fig. 4b, the spacings
of 3.50 Å and 2.35 Å belong to (101) and (112) interplanar distances of anatase TiO . It is also
2
clear from the TEM images (Fig. 4d-4f) that Cu-NPs did not enter into the lattice of TiO , but are
2
well dispersed on its surface. The d-spacings of 1% Cu-Red/TiO and 2% Cu-Red/TiO are 3.52
2
2
Å (101) and 2.07 Å (111) corresponding to the anatase and metallic copper, respectively (Fig. 4e,
f). In case of 1% Cu-Oxi/TiO and 2% Cu-Oxi/TiO photocatalysts, CuO and Cu O
4
2
2
2
nanoparticles have been identified according to interplanar spacings of 2.73 and 2.11, 2.15, 3.05
Å. CuO nano particles in Fig. 4d are of size about 10 nm and may formed by agglomeration of
several smaller CuO nanoparticles.
The XPS spectra of TiO , 2% Cu-Oxi/TiO , 2% Cu-Red/TiO are shown in Fig. 5 revealing the
2
2
2
chemical state of titanium and copper in these catalysts. The binding energy peaks of titanium in
TiO and 2% Cu-Oxi/TiO were observed at overlapping position: 2p3/2 peak at 458.5 eV and
2
2
2
p1/2 peak at 464.5 eV. The 2% Cu-Red/TiO has 2p3/2 peak of Ti at 459.4 eV and 2p1/2 peak
2
at 465.3 eV (Fig. 5a). 2% Cu-Oxi/TiO (Fig. 5b) has a peak at 934.4 eV belonging to 2p3/2 of
2
2
+
Cu in CuO and a broad satellite peak of weak intensity at 943.8 eV while 2p1/2 peak at 954.8
2
+
eV is also a characteristic of Cu species. After reduction of oxi Cu/TiO under hydrogen at 350
2
°
C for 2 h, the position of Cu 2p3/2 peak is shifted with slight modification in shape and appears
as a narrow peak centered at 931.5 eV and disappearance of satellite peak (942.5 eV) indicating
that copper is present in reduced form, probably as metallic copper [37].
1
7

Fig. 4. TEM of a) TiO nanoparticles; b) HR-TEM of TiO nanoparticles; c) SAED pattern of
2
2
TiO nanoparticles with inset EDX; d) 2% Cu-Oxi/TiO with inset EDX; e) and f) 2% Cu-
2
2
Red/TiO2.
1
8

Fig. 5. XPS spectra of TiO and Cu/TiO photocatalysts a) for Ti and b) for copper.
2
2
The copper dispersion measurements of catalysts have been carried out by H -TPR followed by
2
N O passivation and H -TPR. The H consumption peak in H -TPR of CuO in 1% Cu-Oxi/TiO
2
2
2
2
2
is almost half than 2% Cu-Oxi/TiO due to lower loading amount, Fig. 6. Three reduction peaks
2
were observed with 1% Cu-Oxi/TiO at ca. 140, 158 and 195 °C can be attributed to three
2
different size CuO present at the surface of TiO . But 2% Cu-Oxi/TiO has only two peaks at 158
2
2
and 180 °C for two different size CuO nanoparticles. The H -TPR peak temperature is directly
2
proportional to the size of loaded particles i.e. smaller particles reduce at lower temperature [26,
3
8]. Here 135 °C peak can be attributed to fine CuO nanoparticles, 158 °C peak for mid size, 180
°
C peak for larger size and 195 °C peak for largest size CuO nanoparticles. The temperature
programmed N O passivation for both catalysts is shown in SI-Fig. 2 and 3. The lower
2
temperature peak (135 and 158 °C) as well as 195 °C reappeared but the peak at 180 °C does not
attain the good intensity. After passivation 63% and 51% copper dispersion was observed in 1%
Cu-Oxi/TiO and 2% Cu-Oxi/TiO . This indicate the CuO NPs are well dispersed. The 1% Cu-
2
2
Oxi/TiO has comparably better dispersion of copper nano particles than 1% Cu-Oxi/TiO .
2
2
1
9

1
% Cu-Oxi/TiO2
% Cu-Oxi/TiO2
2
50
100
150
200
250
300
350

Temperature ( C)
Fig. 6. H -TPR graph of 1% Cu-Oxi/TiO and 2% Cu-Oxi/TiO .
2
2
2
The fluorescence spectra of TiO , copper loaded catalysts are shown in Fig. 7. The fluorescence
2
spectra of TiO is originated by the recombination of photo generated electrons and holes
2
[
28,39]. The addition of electron/hole trap inhibit the recombination process, here copper act as
an electron trap i.e. reduced number of recombination sites on the TiO surface. Furthermore it is
2
well known that Cu favours the transfer of photogenerated electrons and restrain the
recombination process. Therefore it is reasonable that the fluorescence intensity of copper loaded
TiO is much smaller than pure TiO and resulting in higher catalytic activity [40]. The
2
2
fluorescence spectra follows the order of intensity as TiO > 1% Cu-Oxi/TiO > 2% Cu-
2
2
Oxi/TiO > 2% Cu-Red/TiO > 1% Cu-Red/TiO .
2
2
2
2
0

Fig. 7. PL emission spectra of TiO and Cu/TiO catalysts.
2
2
3
.2
Photocatalytic Activity
Photocatalytic reduction of CO was evaluated in the continuous gas flow reactor under UV
2
illumination at ambient temperature and pressure using 8W mercury lamps serving as light
source. During photocatalytic experiments, CO and H were detected as products. A series of
2
blank experiments was carried out in a preconditioned reactor under the following conditions: a)
using water saturated CO without both catalyst and UV-C irradiation, b) saturated CO and UV-
2
2
C irradiation without catalyst, c) saturated CO and catalyst 2% Cu-Oxi/TiO without irradiation,
2
2
d) in place of CO , water saturated N feed to the reactor having 2% Cu-Oxi/TiO and UV-C
2
2
2
irradiation. No CO conversion/carbon containing product was observed during any of above
2
experiments but and only traces of hydrogen were observed during (d) experiment conditions.
The CO conversion only occurred when photocatalyst and UV irradiation were present together.
2
This demonstrates that CO conversion described below is a true photocatalytic reduction
2
process.
2
1

Fig. 8. Product formation during photocatalytic reduction of CO over 3 h of irradiation:
2
a) CO production with different flow rate over 2% Cu-Oxi/TiO ; b) CO production with different
2
light sources over 2% Cu-Oxi/TiO ; c) CO production over different photocatalysts; d) H
2
2
production over different photocatalysts under UV-C light.
Amongst the catalysts, 2% Cu-Oxi/TiO was taken as a representative for elucidation of the
2
effect of two reaction parameters studied viz. flow rate of CO and wavelength of light source.
2
These parameters were varied in order to get suitable reaction condition for optimum
photocatalytic reduction of CO and also mechanistic insights. For the three different flow rates
2
2
2

of water saturated CO (the catalyst amount, intensity and wavelength of light were fixed) the
2
reaction was carried out for three hours and then the UV-C lamp was switched off. The reaction
products were measured in reactor effluent at every hour during irradiation. As illustrated in Fig.
8
a, the carbon monoxide formation rate steadily increased during the reaction period and reached
its highest value at third hour. CO formation rate at the third hour of reaction improved from 118
-1
-1
-1
to 172 µmolg h on increasing the flow rate of water saturated CO from 10 to 20 mL min .
2
-1 -1
The carbon monoxide formation rate slightly increased to 182 µmolg h on further increase in
-1
CO flow rate to 30 mL min , which suggests that higher flow rate is not an essential condition
2
for good mass transfer inside the photocatalytic reactor. The influence of irradiation wavelength
was investigated using three different light sources of equal electrical power, viz. 8 W visible
-1
light LED, UV-A (350 nm) lamp or UV-C (254 nm) lamp, at 20 mL min flow rate of water
saturated CO and 250 mg catalyst spread in glass reactor. With visible light, no trace of CO was
2
-1 -1
observed but with UV-A and UV-C 26 and 172 µmolg h of CO was observed, respectively.
-1 -1
Here considerable amount of hydrogen formation was observed i.e. 293 µmolg h of H under
2
UV-C only (Fig. 8d).
Other catalysts viz. TiO , 1% Cu-Oxi/TiO , 1% Cu-Red/TiO and 2% Cu-Red/TiO were also
2
2
2
2
-1
tested under the optimized conditions (Fig. 8c and d). Thus 20 mL min flow of water saturated
CO passed over 250 mg catalyst in the preconditioned reactor irradiated with UV-C (254 nm)
2
pen ray lamp. The 1% Cu-Red/TiO has exhibited the highest photocatalytic performance (334
2
-1
-1
-1 -1
µmolg h of CO and 452 µmolg h of H ) while 2% Cu-Red/TiO has production rate 248
2
2
-1
-1
-1 -1
µmolg h for CO and 240 µmolg h for H , which is lower despite increased loading of
2
-1 -1
copper. TiO has given only 160 µmolg h of CO but no hydrogen formation, on the other hand,
2
-1
-1
-1 -1
1
% Cu-Oxi/TiO has given 144 µmolg h for CO and 10 µmolg h for H formation.
2 2
2
3

Trace amount of H was observed during blank test with water saturated N feed to the reactor
2
2
having 2% Cu-Oxi/TiO and UV-C irradiation. But in the presence of CO considerably high H
2
2
2
-1 -1
formation rate i.e. 293 µmolg h was observed. The formation of formic acid during
photocatalytic reduction of CO has been reported in literature [41-44] as well as the
2
photocatalytic splitting of formic acid to H and CO has also been explored [45-46]. Here the
2
2
photocatalytic splitting of formic acid was performed to confirm that hydrogen generation
proceeds as a result of photocatalytic splitting of formic acid/formates formed during the CO2
reduction and which is also a concomitant to CO reduction to CO (SI Fig. 4). It is also reported
2
in literature that hydrogen is produced during photocatalytic splitting of formic acid catalyzed by
copper and TiO [45]. Therefore, formic acid formed during the reaction may provide extra
2
protons to generate hydrogen. Oxidation reactions are different and formic acid oxidation results
in CO in the present case.
2
Fig. 8 demonstrates that CO production rate is much higher under UV-C than under UV-A
irradiation and hydrogen is produced only under UV-C. However, Fig. 2 shows that the
photocatalysts used all have complete absorption of light with wavelengths below 380 nm due to
the TiO band gap value of ca. 3.1 eV. Thus, the difference in activity under UV-C and UV-A
2
light is not due to a difference in the fraction of light absorbed. It could be hypothesized that Cu-
NPs surface could serve as photocatalyst under UV-C irradiation since the energy of 254 nm
quantum of 4.9 eV is large enough to break chemical bonds of formates/bicarbonates weakened
by adsorption on the Cu or CuO surface. To check this hypothesis, Cu-NPs were deposited onto
x
photocatalytically inert silica surface using conditions exactly same as those used for preparation
of Cu/TiO photocatalysts (details are in SI-5).
2
3
.3 Modeling of Cu/TiO and CO Photoreduction
2 2
2
4

Fig. 1 demonstrates that TiO in the Cu/TiO photocatalysts of the present study is present in the
2
2
anatase form. As shown in Fig. 4, Fig. 6 and SI-Fig. S2, S3 Cu-NPs form contact with larger
TiO nanoparticles in the most active reduced Cu/TiO photocatalyst. Both metallic Cu and Cu O
2
2
2
are suggested to be present in the H reduced Cu/TiO catalyst in this and previous studies [20].
2
2
CuO is also formed from the initial Cu and Cu O during the process of CO reduction as a result
2
2
of H O oxidation with O formation [24]. It is well known that the surface of undoped anatase
2
2
nanoparticles is formed mostly by (101) facet due to its lowest surface energy among all the
anatase facets [47]. For this reason the quantum chemical model is based on a Ti O H cluster
1
0
22
4
having structure of (101) facet of anatase. PM6 semiempirical method was used for the modeling
because it is computationally robust and provides good precision when compared to higher level
computational results and experimental data [48-50]. Electrical neutrality of this cluster is
4
+
provided by attaching four hydrogen atoms at the place of lattice Ti ions. Fig. 9A shows this
cluster in the ground singlet electronic state, its enthalpy of formation and total energy.
The process of Cu-NP loading over TiO nanoparticles is represented here by attaching Cu
2
4
clusters to the (101) surface of the above (TiO ) (H O) cluster as revealed in Fig. 9. Cu clusters
2
10
2
2
quantum mechanical modeling have been reported in several publications, for example, in
47,51]. It was shown that a planar geometry is more stable for the Cu cluster in vacuum. We
[
4
identified two such planar geometries both having ground singlet state in agreement with the
literature data that Cu clusters are singlet for even x and doublet for odd x [51]. In the first
x
geometry shown in Fig. 9B, the Cu has a parallelogram structure with Cu-Cu bond length 2.29,
4
2
.54 and 3.16 Å. However, the most stable Cu cluster has the shape of an isosceles triangle (Cu-
4
Cu bonds 2.29 Å (x2) and 2.61 Å ) with one external attached atom (Cu-Cu bond 2.29 Å). As
seen in Fig. 9C, this structure is about 10 kcal/mol more stable than the parallelogram one.
2
5

Fig. 9. Scheme of copper nanoparticle deposition on TiO nanocluster. Atoms designations: ● -
2
oxygen, ● - titanium, ● - copper, ● - hydrogen.
TiO facet (101) has two types of atoms exposed - oxygen atoms O2C coordinated to two
2
neighbor Ti atoms and Ti5C sites coordinated to five oxygen atoms. Cu cluster tends to be
4
attached to O2C sites only while keeping the distance Ti-Cu above 3 Å. Three modes of Cu4
attachment have been identified. In the first one, Cu forms a planar quadrangle attached to O
4
2C
via a single Cu atom to form the 2.3 Å Cu-O bond. In the second mode shown in Fig. 9 D, Cu4
has a tetrahedron-like shape with two Cu atoms attached to the same O2C via 2.3 Å bonds. In the
third geometry shown in Fig. 9E, two Cu atoms of the distorted tetrahedron are attached to two
neighbor O2C forming 2.1 Å bonds. The enthalpy of Cu adsorption was calculated with the
4
following equation

Hads =  H(Cu Ti O H ) -  H(Cu ) -  H(Ti O H ).
f
4
10 22
4
f
4
f
10 22
4
2
6

It turned out that the first planar mode of adsorption is stable with Hads = -9 kcal/mol, the
second mode of adsorption is unstable with Hads = +13.07 kcal/mol while the third mode is very
stable with Hads = -26.01 kcal/mol. The tetrahedral Cu cluster bound to two O sites over
4
2C
(101) face have also been reported in [52] as the most stable structure while a planar adsorption
geometry less stable. Since Cu cluster tended to form multiple Cu-O coordination bonds it is
4
2C
naturally concluded that the real Cu nanoparticles would also tend to form the same type of
bonds over the prevalent (101) anatase surface. In the TiO -Cu interfacial region of the Cu-
2
Red/TiO photocatalysts, it is thus expected to meet strong Cu-O bonding while leaving the
2
2C
adjacent Ti sites unoccupied.
5
C
Fig. 10. Photocatalytic cycle of CO reduction to CO at the Cu/TiO interface. Atoms
2
2
designations: ● - oxygen, ● - titanium, ● - copper, ● - hydrogen, ● - carbon.
2
7

The process of CO photocatalytic reduction starts with CO molecule adsorption on Cu/TiO
2
2
2
photocatalyst surface. The adsorption can take place over TiO surface, Cu surface or at the Cu-
2
TiO interface. Over clean (101) TiO anatase surface, two modes of CO adsorption have been
2
2
2
identified. In the first one, CO molecule is coordinated to Ti5C with oxygen atom forming COTi
2
angle 131° and slightly deviating from linearity 169° bent adsorbed CO molecule. In the second
2
mode, CO is coordinated via carbon atom to O to form a carbonate-like structure with bonds
2
2C
Ti-O 2.3 Å, C-O2C 1.6 Å and OCO angle 143°. Enthalpy of CO adsorption for these two modes
2
is -9.45 and -6.17 kcal/mol giving evidence for the predominance of linear adsorption with
oxygen coordination over unmodified anatase (101) face.
The process of CO adsorption at the Cu-TiO interface is depicted in Fig. 10 as transition from
2
2
structure A to structure B. Here the most stable adsorption mode is in the carbonate form (Fig.
0B) with carbon atom coordinated to O2C and oxygen to Cu. Bond lengths are C-O 1.51 Å, Cu-
1
O 2.07 Å and the OCO angle is 134°. Enthalpy of CO adsorption in this form is -29.54
2
kcal/mol. Adsorption over Cu cluster loaded onto TiO was found to be weakly exothermic
4
2

Hads = -1.81 kcal/mol. Consideration of the relative energetics of CO adsorption over different
2
surfaces of Cu/TiO catalyst leads us to conclusion that Cu-TiO interface provides the strongest
2
2
mode of CO adsorption which is realized in the carbonate form.
2
After UV-A or UV-C quantum absorption in the TiO nanoparticle, photogenerated electrons and
2
holes can recombine or migrate to the surface where they take part in CO reduction and water
2
oxidation, respectively. The process of CO (ads) reduction in humidified conditions consists of
2
attachment of photogenerated electron and then a proton. These two steps result in structure C in
Fig. 10. If we consider the reaction B → C as attachment of hydrogen atom H˙ to the adsorbed
CO then this reaction has  H = -44.2 kcal/mol, the energy demanding step being formation of
2
r
2
8

highly energetic hydrogen atom with  H = +52.10 kcal/mol. However, the single electron
f
reduction of structure B is predicted to be highly exothermic in vacuum. This agrees with the
recent finding on spontaneous CO to CO reduction in the dark over reduced Cu/TiO catalysts
2
2
[
24].
The next step is dissociation of the HOCO(ads) to form separately adsorbed hydroxyl group and
carbon monoxide. The hydroxyl group is transferred to the neighbor Ti5C while CO remains
bound to the same O2C and Cu atoms as shown in Fig. 10D. The C → D reaction is slightly
exothermic with  H = -2.63 kcal/mol. The bidentate binding of CO with oxygen coordination to
r
Cu is highly preferred here: the structure E with carbon atom coordination to Cu is by 55.12
kcal/mol less stable.
Desorption of CO can be considered here as a two-step process D → E → F with the reaction
enthalpy +55.12 and +20.21 kcal/mol, or as a single-step D → F with  H = +75.33 kcal/mol.
r
The difficulty of CO desorption can explain very strong positive influence of UV-C radiation on
the reaction rate of CO to CO photoreduction. Indeed, such highly energetic quanta are not
2
needed for creation of electron-hole pair but can be absorbed by surface bound carbonyl CO
molecules resulting in their excitation, dissociation of the carbonyls and desorption. Photolysis of
surface carbonyls can result in a different arrangement of Cu atoms as shown in Fig. 10F
preventing back-binding of CO molecules.
The final step of the catalytic cycle is the second-electron reduction of the hydroxyl
group to form water and restored (101) surface ready for the next turn of CO reduction as shown
2
by F → A transformation in Fig. 10. The enthalpy change in this step is -72.66 kcal/mol if
enthalpy of formation of electron and proton are taken equal to that of hydrogen atom.
2
9