Embedded CuO nanoparticles@TiO2-nanotube arrays for photoelectrocatalytic reduction of CO2 to methanol

Article abstract of DOI:10.1016/j.electacta.2018.07.072

Herein, we report a simple approach to synthesize high-performance and structurally stable CuO nanoparticles catalyst embedded in TiO₂ nanotube arrays for CO₂ reduction. To summarize, anodic TiO₂ nanotube arrays (TNTs) were electrochemical reductive doped in 1 mol/L (NH₄)₂SO₄ solution to form an activated surface. Then CuO nanoparticles were successfully filled into the pores pace of TNTs by electrodeposition and heat treatment. The effects of Ti(Ⅲ) reduction doping were discussed by means of electrochemical impedance spectroscopy (EIS) and X ray photoelectron spectroscopy (XPS). Results show that partial Ti(Ⅳ) in TNTs can be reduced to Ti(Ⅲ) by electrochemical reduction, which leads to a significant improvement in TNTs surfactivity and then benefits the deposition of Cu nanoparticles to form a stable embedded structure. As a consequence, the composite electrodes showed higher photoelectrocatalytic performance for CO₂ reduction. The maximum current of CuO-TNTs composite electrodes is up to ?1.37 mA/cm² at ?0.5 V and high selectivity for methanol synthesis is also obtained in this case. At the same time, the amount of methanol produced by the CuO/self-doped TNTs composite electrode is about 15% higher than that by the CuO/TNTs electrode without reductive doping.

Full text of DOI:10.1016/j.electacta.2018.07.072

Accepted Manuscript
Embedded CuO nanoparticles@TiO -nanotube arrays for photoelectrocatalytic
2
reduction of CO to methanol
2
Liqiang Zhang, Huazhen Cao, Qiuyuan Pen, Liankui Wu, Guangya Hou, Yiping Tang,
Guoqu Zheng
PII:
S0013-4686(18)31573-1
10.1016/j.electacta.2018.07.072
EA 32269
DOI:
Reference:
To appear in: Electrochimica Acta
Received Date: 19 June 2018
Revised Date: 9 July 2018
Accepted Date: 12 July 2018
Please cite this article as: L. Zhang, H. Cao, Q. Pen, L. Wu, G. Hou, Y. Tang, G. Zheng, Embedded
CuO nanoparticles@TiO -nanotube arrays for photoelectrocatalytic reduction of CO to methanol,
2
2
Electrochimica Acta (2018), doi: 10.1016/j.electacta.2018.07.072.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Embedded CuO nanoparticles@TiO2-nanotube arrays for
photoelectrocatalytic reduction of CO2 to methanol
Liqiang Zhang, Huazhen Cao*, Qiuyuan Pen, Liankui Wu, Guangya Hou, Yiping
Tang,
Guoqu Zheng*
(College of Materials Science and Engineering, Zhejiang University of Technology,
Hangzhou 310014, China)
ABSTRACT
Herein, we report a simple approach to synthesize high-performance and structurally
stable CuO nanoparticles catalyst embedded in TiO2 nanotube arrays for CO2
reduction. To summarize, anodic TiO2 nanotube arrays (TNTs) were electrochemical
reductive doped in 1 mol/L (NH4)2SO4 solution to form an activated surface. Then
CuO nanoparticles were successfully filled into the pores pace of TNTs by
electrodeposition and heat treatment. The effects of Ti(ꢀ) reduction doping were
discussed by means of electrochemical impedance spectroscopy (EIS) and X ray
photoelectron spectroscopy (XPS). Results show that partial Ti(ꢁ) in TNTs can be
reduced to Ti(ꢀ) by electrochemical reduction, which leads to an significant
improvement in TNTs surfactivity and then benefits the deposition of Cu
nanoparticles to form a stable embedded structure. As a consequence, the composite
electrodes showed higher photoelectrocatalytic performance for CO2 reduction. The
maximum current of CuO-TNTs composite electrodes is up to -1.37 mA/cm2 at -0.5 V
and high selectivity for methanol synthesis is also obtained in this case. At the same
time, the amount of methanol produced by the CuO/self-doped TNTs composite
electrode is about 15% higher than that by the CuO/TNTs electrode without reductive
doping.
Keywords: CuO catalyst, TNTs reductive doping, Photoelectrocatalysis, CO2

ACCEPTED MANUSCRIPT
reduction
1 Introduction
Global warming and energy shortage are becoming two big problems in the
21th century. The problems caused by the increasing CO2 concentration in the
atmosphere are damaging our living environment and bringing negative effects to
sustainable development. Many researchers have proposed the use of photocatalytic
reduction of CO2 to organics (such as methanol, formaldehyde, methane, etc.) to solve
these problems. Cu and its oxides [1-4] have attracted extensive research interest in
photocatalytic or photoelectrocatalytic reduction of CO2 due to the narrow band gap,
relative negative conduct band, low cost and non-toxicity. For example, CuO with
band gap of 1.3~1.6 eV shows high visible light response. Its relative negative
conduct band is about -0.78 V vs. NHE, far below the reduction potential of CO2 to
CH3OH (-0.38 V vs. NHE) and CO2 to HCHO (-0.48 V vs. NHE) [5, 6], which means
a strong ability for CO2 reduction.
Some studies have reported the excellent photocatalytic activity of CuO
nanoparticles [7, 8]. As we known, the agglomeration of CuO nanoparticles and the
charge recombination are the important factors leading to the decrease of their
photocatalytic activities. In order to solve these issues, various strategies have been
tried, for example, the use of dispersant [7] and constructing heterojunction
semiconductors system to promote electron-hole separation [9]. TNTs is a favorable
substrate in nanometer material synthesis [10-15] owing to its distinctive
nanostructure and high chemical stability [16, 17]. Furthermore, TNTs in itself is a
kind of photocatalyst responding to only UV light because of its broad bandgap [18].
Thus, the combination of CuO with TNTs may bring out some expected results, i.e.
the visible light response and the p-n junction arising from the p type CuO and the n
type TiO2, which can efficiently promoted the separation of photoelectrons from
vacancies.
However, CuO nanoparticles are easy to accumulate on the surface of TNTs,

ACCEPTED MANUSCRIPT
and few can be deposited into the pore space of TNTs, which is mainly caused by the
poor conductivity and the inert surface of TiO2 nanotubes [19]. The loose crumb
structure usually leads to a rapid depletion of CuO nanoparticles, accompanied by fast
decay in photocurrent and low efficiency. In the prior studies [20], it has been
discovered that partial Ti(ꢁ) can be reduced to Ti(ꢀ) by reductive doping, and the
conductivity and surficial characteristics of TNTs can be improved remarkably, thus
facilitating the deposition of nanoparticles. Based on this, we synthesized CuO-TNTs
composite electrode with a stable embedded structure via using reductive doped TNTs
as substrate, in order to obtain high-performance and structurally stable CuO
composite catalyst.
2 Experimental
2.1 Preparation of CuO/TNTs.
TNTs are prepared by anodic oxidation on Ti sheets (25 mm×15 mm×1
mm, >99.5%) in 1% HF aqueous solution at constant potential (20V) and room
temperature (20°C). Before anodization, the Ti sheets were polished with
metallographic sandpaper and ultrasonically degreased in ethanol, followed by
chemical polishing in solution with 50 g/L CrO3 and 3% HF at 50°C. All TNTs were
heat treated under 450°C for two hours in a tube furnace with heating and cooling rate
of 5°C per minute, by which the amorphous TNTs can be transformed into anatase [21,
22].
Reductive doping of TNTs was conducted in 1 mol/L (NH4)2SO4 solution at
different applied potentials with different reduction times using CHI660C
electrochemical workstation. A conventional three-electrode system was used by
employing the TNTs, Pt sheet and a saturated calomel electrode (SCE) as working
electrode, counter electrode and reference electrode, respectively. Cu was
electrodeposited on the reductive doped TNTs in 1 mol/L CuSO4 via pulse mode with
a cathodic pulse (-0.07A, 0.01s), an anodic pulse (+0.07A, 0.002s) and rest time (0A,
1s). Finally, all samples were calcined at 450°C for 1 h.

ACCEPTED MANUSCRIPT
2.2 Characterization
Cyclic voltammetry curves and electrochemical impedance spectroscopy were
measured on CHI660C electrochemical workstation using platinum sheet and
saturated calomel electrode (SCE) as auxiliary electrode and referring electrode,
respectively. The frequency range of 100 kHz to 10 mHz and the modulation
amplitude of 10 mV were employed for impedance measurements. Microstructural
observation and elemental analysis were performed by TEM (Tecnai G2 F30) and
SEM (ZEISS SUPRA55) equipped with the GENENIS-4000 energy dispersive
spectrometry (EDS). XRD patterns of TNTs and CuO/TNTs were recorded on a
Panalytical X'Pert PRO equipped with Cu Kα radiation (λ=0.154056 nm) at 40 kV
and 40 mA. And the valence states of Ti in TNTs were analyzed by XPS (Kratos Axis
ultra DLD, Al Kα X-ray source, hν=1486.6 eV).
2.3 Photoelectrocatalytic reaction
The photoelectrocatalytic reduction of CO2 was carried out in a homemade 100
ml gas tight dual-chamber cell with 0.1 mol/L NaHCO3 solution, in which the
CuO/TNTs electrode with an exposed area of 1 cm×1 cm and a Pt foil were served as
photocathode and counter electrode. The chamber was irradiated from the side by a
250 W xenon lamp and the light intensity reaching the surface of sample was about
100 mW/cm2. CO2 was kept bubbling into the solution to remove oxygen completely
and saturate CO2 in solution. The photocurrent curves were recorded at −0.5 V (vs
SCE) with an interval of 60 s for light on/off. The total yield of methanol and
formaldehyde in the liquid phase from the photoelectrocatalytic reduction of CO2 in
the period of 5h was detected by headspace gas chromatograph (Fuli GC9720)
equipped with a flame ionization detector (GC-FID) and capillary column (30m, inner
diameter 0.25mm, DB-23) . The column and detector temperatures were kept at 60
and 180 °C, respectively. The carrier gas of high-purity N2 was flowing at a rate of 30
mL/min. The amount of formaldehyde was detected by UV−vis spectroscopy (HACH
DR6000). In detail, 10 mL liquid product was directly mixed with 2.5 ml of color
agent which consists of 50 g ammonium acetate, 6 ml acetic acid and 1ml

ACCEPTED MANUSCRIPT
acetylacetone in 100 mL deionized water. After mixing, the final solution was heated
at 50°C for 20 min and analyzed by UV−vis spectroscopy.
3. Results and discussion
3.1 Ti(ꢀ) reductive doping
Redox behavior of TNTs in (NH4)2SO4 aqueous solution is studied and the
characteristic CV curve is shown in Fig. 1. A cathodic peak is clearly seen at -1.46 V
during negative potential scanning, which denotes to the reduction of Ti((ꢁ) to Ti(ꢀ)
[20]. On the reverse scanning, an oxidation peak corresponding to the reduction of
Ti(ꢀ) appears at -1.30 V. The CV curve indicates that the self-doping of TNTs can be
conducted by electrochemical reduction under certain potentials.
Naturally, the reductive doping potential is a major factor to be considered. Low
overpotential is insufficient to drive the reduction of Ti(ꢁ) and under too negative
potential serve hydrogen evolution will occur. In order to reveal the exact influence of
doping potential on the surface reactivity of TNTs for Cu electrodeposition, a series of
electrochemical impedance curves are measured for various reductive-doped TNTs, as
shown in Fig. 2. It is found that all curves show a slightly depressed semicircular
shape. The different radius indicates different surfactivity in medium. It is obvious
that the reductive doped TNTs is more active than the sample without reductive
doping, as evidenced by the shrunken capacitive loops. An equivalent circuit model
Rs(RctQ) is employed to analyze the EIS data. In this circuit, Rs stands for electrolyte
resistance, Rct parallel accounts for the charge transfer resistance of Cu reduction on
TNTs and Q is constant phase angle component. Table. 1 shows the fitting results by
using the equivalent circuit model Rs(RctQ). It is found that the charge transfer
resistance is as high as 10.8 kꢀ cm2 for the undoped TNTs, which reduces obviously
when the TNTs is reductive doped. Moreover, along with the reductive potential
turning negative the charge transfer resistance decreases and then reaches a minimum
at potential of -1.5 V, after which it changes little.
In order to further illustrate the exact effects of reductive doping, XPS analysis is
adopted to reveal the variation of Ti valence state in TNTs. From the XPS patterns in

ACCEPTED MANUSCRIPT
Fig. 3, it is noted that the characteristic peak of Ti2p XPS for undoped TNTs appears
at 458.8 eV [21, 23], which shifts slightly toward low binding energy after reductive
doping at -1.2, -1.5 and -1.7 V. The negative shift of Ti2p peak is an evidence of Ti(ꢀ)
existence [24-26]. By comparison, it is also found that the doping amount of Ti(ꢀ)
reaches saturation at -1.5 V and the further decreasing of doping potential is bootless.
This result agrees well with the electrochemical impedance analysis.
3.2 Characteristic of CuO/TNTs
Fig. 4 shows the XRD patterns of TNTs and CuO/TNTs composite electrode, in
which the TNTs support is reductive doped at -1.5 V. The diffraction peaks of Ti
substrate is clearly seen as the CuO/TNTs film is only several hundred nanometers
thick. The weak diffraction peaks at 25.28° denotes the presence of TNTs (JCPDS
00-001–0562) [27], and the peaks emerging at 35.59° and 38.78° are attributed to (0 0
2) and (1 1 1) reflections of face-centered cube CuO. There is no obvious Cu2O
diffraction peak, indicating that the heat treatment at 450°C enables a complete
transformation of the electrodeposited Cu to CuO. It must be noted that the Ti(ꢀ) will
also be oxidized to Ti(ꢁ) during the heat treatment because of the unstable surface
doping [28].
Fig. 5 shows the top and cross-section views of CuO nanoparticles loaded on the
TNTs support with or without reductive doping. It can be seen that the TiO2 nanotubes
exhibit an average diameter of 80 nm with tube length of 300~400 nm. The deposits,
i.e. CuO nanoparticles, are clearly seen while their filling position are much different.
For the case of undoped TNTs (Fig. 5a, b), lots of large-grained CuO particles
accumulate on the surface of TNTs and only few are filled into the pore space of
TNTs, so many nanotubes are empty. However, when using the TNTs reductive doped
at -1.5 V as support (Fig. 5c, d), it is readily to find that numerous CuO nanoparticles
are deposited in the pore space of TNTs and their particle size becomes smaller.
Obviously, the deposition of CuO on TNTs support is facilitated by the reductive
doping of TNTs. A CuO/TNTs composite electrode with an embedded structure is
then synthesized, which seems more stable than the one with CuO particles

ACCEPTED MANUSCRIPT
accumulating on the surface of TNTs due to the space limitation by TiO2 nanotubes.
The corresponding TEM images of CuO/TNTs with reductive doping are shown
in Fig. 6a. Similarly to the above SEM observations, the TEM image also
demonstrates that the CuO nanoparticles are well-deposited in the pose space of TNTs,
which locate at not only the inner side of nanotubes but also the interstice between
nanotubes (Fig. 6a). In addition, most CuO nanoparticles grow along the tubes wall
and so some tubes are empty at the center, i.e. forming a hollow structure. In Fig. 6b,
the HRTEM image exhibits two different set of atomic lattice fringes, and they match
well with the (1 0 3) crystallographic plane of anatase TiO2 and (-1 1 1)
crystallographic plane of CuO. This result is consistent with the XRD analysis in Fig.
4. By contrast, the deposition of Cu nanoparticles in undoped TNTs is unsatisfactory
(Fig. 6c).
Based on the above analysis, a schematic illustration of the growth of Cu
nanoparticles on different TNTs is presented in Fig. 7. By reductive doping treatment,
partial Ti(ꢁ) can be reduced to Ti(ꢀ) at the presence of H+ (ꢀꢁꢂꢃ + ꢄꢅ + ꢆꢃ →
ꢀꢁꢇꢃꢆꢃ) [19]. Both the bottom and the wall of TiO2 tubes is then activated due to
existence of these active sites. The enhancement in surfactivity is bound to bring
convenience for the deposition of Cu nanoparticles. As a result, Cu nanoparticles will
grow from the bottom and the tube wall, leading to form an embedded structure.
However, for the case of bare TNTs without any treatment, it is difficult for Cu
nanoparticles to deposit inside the nanotubes due to its inert state. In this way,
abundant of Cu nanoparticles can only grow on the surface of TNTs and form loose
crumb structure.
3.3 Photoelectrocatalytic activities of CuO/TNTs
The linear sweep voltammetry of CuO/TNTs in CO2/N2 saturated 0.1 M Na2SO4
solution was shown in Fig. 8. In the medium without carbon source (N2-saturated), the
cathodic current increases sharply when the scanning potential reaches -0.6 V, which
denotes hydrogen evolution. It is found that the cathodic current in CO2-saturated
medium is much higher than that in N2-saturated medium, indicating that the electrode

ACCEPTED MANUSCRIPT
has excellent electrocatalytic/photoelectrocatalytic activity for CO2 reduction.
Fig. 9a shows a series of transient photocurrent curves of CuO/TNTs electrodes
using different TNTs support, in order to understand the effects of reductive doping
potential on their photoelectrocatalytic properties for CO2 reduction. For comparison,
the transient photocurrent curve of the sample prepared on titanium plate is also
presented. It can be seen that TNTs play a significant role in photoelectrocatalysis, the
photocurrent on a CuO/TNTs electrode is much higher than that on a CuO/Ti
electrode. From these curves, the effects of reductive doping potential of TNTs on
electrode’s photoelectrocatalytic property are evident. The initial photocurrent of
composite electrode is only around -1.21 mA/cm2 when using undoped TNTs as
support. However, for the case of using doped TNTs as support, the initial
photocurrent rises obviously, which is about -1.28 mA/cm2, -1.37 mA/cm2 and -1.37
mA/cm2 respectively at doping potential of -1.2 V, -1.5 V and -1.7 V. Also, it is found
that when the doping potential of TNTs is not negative enough (i.e. -1.2 V), the
effects of reductive doping is not obvious. With the increasing of doping potentials,
the photocurrent rises and reaches the maximum at potential of -1.5 V. Obviously,
Overnegative doping potential won’t improve the performance of CuO/TNTs
electrode any more, which is in agreement with the results in EIS and XPS analysis
(Fig. 2 and Fig. 3).
The influence of reductive doping time at same reduction potential is also
investigated and the corresponding transient photocurrent curves of CO2 reduction are
presented in Fig. 9b. There is an obvious increasing tendency in photocurrent with the
extension of doping time. When the doping time of TNTs is 5 s, an obvious current
growth is observed, which indicates a fast transformation of Ti(ꢁ) to Ti(ꢀ) during
doping process. And it is found that the photocurrent reaches the highest at reductive
doping time of 10 s after which the current keeps constant, that is to say, the reductive
doping of TNTs reaches saturation state quickly.
Furthermore, it is worth noting that whether Ti(ꢀ) is doped or not, there is
severe current fluctuation between light on and off, up to 1mA, which indicates a
strong photoresponse of CuO/TNTs in visible light. In addition, the current of

ACCEPTED MANUSCRIPT
as-prepared CuO/TNTs can reach a high level at the moment light on, which is
contributed to the fast separation and transmission of electron-hole pairs.
3.4 Photoelectrocatalytic reduction of CO2
Photoelectrocatalytic properties of CuO/TNTs electrodes toward CO2 reduction
is investigated at -0.5 V under visible light. The product contains methanol and
formaldehyde in the liquid phase and very few methane is detected in the gas phase.
The methanol yield by using the synthesized CuO/TNTs electrode is displayed in Fig.
10. It is found that the methanol yields of the two samples synthesized on different
TNTs vary greatly with time. For instance, after one hour of reaction, the methanol
yield on CuO/self-doped TNTs and CuO/undoped TNTs are 8.3 ꢁmol/cm2 and 7.6
ꢁmol/cm2, respectively. In spite of the subtle difference in the initial stage, however,
after two hours of reaction, the amount of methanol produced by a CuO/self-doped
TNTs electrode is about 15% higher than that by a CuO/undoped TNTs electrode. And
after five hours of reaction, the amount of methanol reaches 30 ꢁmol/cm2 on
CuO/self-doped TNTs electrode, far more than that with a CuO/undoped TNTs
electrode (26 ꢁmol/cm2). The enhancement of methanol yield after long-term reaction
is closely related with the higher activity and the excellent stability of CuO/self-doped
TNTs electrode.
Effect of applied potentials on the products selectivity was also studied and
shown in Fig. 11. It can be seen clearly that in the range of -0.2 V to -0.5 V, the yield
of methanol on CuO/self-doped TNTs electrode increases with the negative shift of
potentials, while the yield of formaldehyde decreases gradually. However, when the
applied potential reaches -0.6 V, the outputs of both methanol and formaldehyde will
decrease. Obviously, formaldehyde is easier to produce during CO2 reduction at low
applied potential (-0.2 V) [29], while methanol is incline to generate at more negative
potentials (>-0.3 V). However, when the applied potential is too negative (>-0.6 V),
the hydrogen evolution reaction began to take place [2, 30], as shown in Fig. 8. As a
result, the competition between CO2 reduction and hydrogen evolution reaction
becomes serious, which finally lead to the decrease of catalytic efficiency for CO2

ACCEPTED MANUSCRIPT
reduction. i.e. the reduction of methanol and formaldehyde yields.
3.5 The reaction mechanism
Fig. 12 shows the diagram of p-n junction of CuO-TNTs. P-type CuO and n-type
TiO2 are combined to form Z‑Scheme p-n junction [31, 32]. Some photo generated
electrons in TiO2 with a lower conduction band could recombine with the holes in
CuO as the valence band of CuO is close to the conduction band of TiO2. More
powerful photo generated electrons and holes can be retained on different counterparts,
the isolated powerful oxidative holes and reductive electrons generated on TiO2 and
CuO respectively result in a strong photoelectrocatalytic ability [32], as shown in Fig.
9. A possible reaction mechanism is that CO2 can be reduced to methanol and
formaldehyde during the reaction because of the powerful reductive electrons
generated by CuO, and formaldehyde will be further reduced to methanol at a
relatively negative potential.
4 conclusion
In conclusion, CuO/TNTs composite electrodes were successfully synthesized by
using reductive doped TNTs as support. Various tests show that the surfactivity of
TNTs doped with Ti(ꢀ) is greatly improved so that Cu nanoparticles can be better
deposited inside the TiO2 nanotubes. The optimum condition for TNTs doping is to
conduct at potential of -1.5 V for 60 s. The corresponding CuO/reductive doped TNTs
electrode shows a strong photocurrent of -1.37 mA/cm2, which means a high
photoelectrocatalytic reduction ability for CO2 reduction. The appropriate potential
for methanol synthesis is around -0.5 V due to the higher yield and the better
selectivity.
Acknowledgements
This work was supported by the Natural Science Foundation of Zhejiang
Province (No. LY17B030009 and No. LQ16E020002).

ACCEPTED MANUSCRIPT
References
[1] G. Ghadimkhani, N.R. de Tacconi, W. Chanmanee, C. Janaky, K. Rajeshwar, Efficient solar
photoelectrosynthesis of methanol from carbon dioxide using hybrid CuO-Cu2O
semiconductor nanorod arrays, Chemical Communications, 49 (2013) 1297-1299.
[2] F.D.B. Juliana, A.D.S. Alexander, A.J. Cavalheiro, M.V. Boldrin Zanoni, Evaluation of the
parameters affecting the photoelectrocatalytic reduction of CO2 to CH3OH at Cu/Cu2O
electrode, International Journal of Electrochemical Science, 9 (2014) 5961-5973.
[3] B. Fang, Y. Xing, A. Bonakdarpour, S. Zhang, D.P. Wilkinson, Hierarchical CuO–TiO2 hollow
microspheres for highly efficient photodriven reduction of CO2 to CH4, ACS Sustainable
Chemistry & Engineering, 3 (2015) 2381-2388.
[4] M. Zhang, J. Cheng, X. Xuan, J. Zhou, K. Cen, Pt/graphene aerogel deposited in Cu foam as a
3D binder-free cathode for CO2 reduction into liquid chemicals in a TiO2 photoanode-driven
photoelectrochemical cell, Chemical Engineering Journal, 322 (2017) 22-32.
[5] S.N. Habisreutinger, L. Schmidt-Mende, J.K. Stolarczyk, Photocatalytic reduction of CO2 on
TiO2 and other semiconductors, Angewandte Chemie International Edition, 52 (2013)
7372-7408.
[6] K. Hashimoto, H. Irie, A. Fujishima, TiO2 photocatalysis: a historical overview and future
prospects, Japanese Journal of Applied Physics, 44 (2005) 8269-8285.
[7] S. Sonia, S. Poongodi, P.S. Kumar, D. Mangalaraj, N. Ponpandian, C. Viswanathan,
Hydrothermal synthesis of highly stable CuO nanostructures for efficient photocatalytic
degradation of organic dyes, Materials Science in Semiconductor Processing, 30 (2015)
585-591.
[8] D. Jiang, J. Xue, L. Wu, W. Zhou, Y. Zhang, X. Li, Photocatalytic performance enhancement
of CuO/Cu2O heterostructures for photodegradation of organic dyes: Effects of CuO
morphology, Applied Catalysis B: Environmental, 211 (2017) 199-204.
[9] V. Scuderi, G. Amiard, R. Sanz, S. Boninelli, G. Impellizzeri, V. Privitera, TiO2 coated CuO
nanowire array: Ultrathin p–n heterojunction to modulate cationic/anionic dye
photo-degradation in water, Applied Surface Science, 416 (2017) 885-890.
[10] J.S. Zhong, Q.Y. Wang, Y.F. Yu, Solvothermal preparation of Ag nanoparticles sensitized TiO2
nanotube arrays with enhanced photoelectrochemical performance, Journal of Alloys and
Compounds, 620 (2015) 168-171.
[11] G.S. Pozan, M. Isleyen, S. Gokcen, Transition metal coated TiO2 nanoparticles: Synthesis,
characterization and their photocatalytic activity, Applied Catalysis B: Environmental,
140-141 (2013) 537-545.
[12] K. Zbudniewek, J. Góralski, J. Rynkowski, Studies on TiO2/SiO2 and Pd/TiO2/SiO2 catalysts
in photoreduction of CO2 with H2O to methanol, Russian Journal of Physical Chemistry A,
86 (2012) 2057-2062.
[13] R. Fagan, D.W. Synnott, D.E. McCormack, S.C. Pillai, An effective method for the
preparation of high temperature stable anatase TiO2 photocatalysts, Applied Surface Science,
371 (2016) 447-452.
[14] S. Kuang, L. Yang, S. Luo, Q. Cai, Fabrication, characterization and photoelectrochemical
properties of Fe2O3 modified TiO2 nanotube arrays, Applied Surface Science, 255 (2009)

ACCEPTED MANUSCRIPT
7385-7388.
[15] C.W. Lai, S. Sreekantan, P.S. E, W. Krengvirat, Preparation and photoelectrochemical
characterization of WO3-loaded TiO2 nanotube arrays via radio frequency sputtering,
Electrochimica Acta, 77 (2012) 128-136.
[16] X. Yan, X. Chen, Titanium Dioxide Nanomaterials, 2015.
[17] M. Hepel, I. Kumarihamy, C.J. Zhong, Nanoporous TiO2-supported bimetallic catalysts for
methanol oxidation in acidic media, Electrochemistry Communications, 8 (2006) 1439-1444.
[18] P. Roy, S. Berger, P. Schmuki, TiO2 nanotubes: synthesis and applications, Angewandte
Chemie, 50 (2011) 2904-2939.
[19] J.M. Macak, B.G. Gong, M. Hueppe, P. Schmuki, Filling of TiO2 nanotubes by self-doping
and electrodeposition, Advanced Materials, 19 (2007) 3027-3031.
[20] H. Cao, K. Huang, L. Wu, G. Hou, Y. Tang, G. Zheng, Enhanced catalytic performance of
Pt/TNTs composite electrode by reductive doping of TNTs, Applied Surface Science, 364
(2016) 257-263.
[21] H. Miao, X. Hu, J. Fan, C. Li, Q. Sun, Y. Hao, G. Zhang, J. Bai, X. Hou, Hydrothermal
synthesis of TiO2 nanostructure films and their photoelectrochemical properties, Applied
Surface Science, 358 (2015) 418-424.
[22] M. Xiao, K.P. Musselman, W.W. Duley, N.Y. Zhou, Resistive switching memory of TiO2
nanowire networks grown on Ti foil by a single hydrothermal method, Nano-Micro Letters, 9
(2016).
[23] J. Jiang, H. Zhou, F. Zhang, T. Fan, D. Zhang, Hydrothermal synthesis of core–shell TiO2 to
enhance the photocatalytic hydrogen evolution, Applied Surface Science, 368 (2016)
309-315.
[24] H. Yin, X. Wang, L. Wang, Q. N, H. Zhao, Self-doped TiO2 hierarchical hollow spheres with
enhanced visible-light photocatalytic activity, Journal of Alloys and Compounds, 640 (2015)
68-74.
[25] W.-D. Zhu, C.-W. Wang, J.-B. Chen, Y. Li, J. Wang, Enhanced field emission from Ti3+
self-doped TiO2 nanotube arrays synthesized by a facile cathodic reduction process, Applied
Surface Science, 301 (2014) 525-529.
[26] W. Liao, J. Yang, H. Zhou, M. Murugananthan, Y. Zhang, Electrochemically self-doped TiO2
nanotube arrays for efficient visible light photoelectrocatalytic degradation of contaminants,
Electrochimica Acta, 136 (2014) 310-317.
[27] H. Wu, J. Fan, E. Liu, X. Hu, Y. Ma, X. Fan, Y. Li, C. Tang, Facile hydrothermal synthesis of
TiO2 nanospindles-reduced graphene oxide composite with a enhanced photocatalytic activity,
Journal of Alloys and Compounds, 623 (2015) 298-303.
[28] M. Wen, S. Zhang, W. Dai, G. Li, D. Zhang, In situ synthesis of Ti3+ self-doped mesoporous
TiO2 as a durable photocatalyst for environmental remediation, Chinese Journal of Catalysis,
36 (2015) 2095-2102.
[29] Y. Ji, Y. Luo, Theoretical study on the mechanism of photoreduction of CO2 to CH4 on the
anatase TiO2(101) Surface, ACS Catalysis, 6 (2016) 2018-2025.
[30] P. Li, J. Zhang, H. Wang, H. Jing, J. Xu, X. Sui, H. Hu, H. Yin, The photoelectric catalytic
reduction of CO2 to methanol on CdSeTe NSs/TiO2 NTs, Catalysis Science &Technology, 4
(2014) 1070-1077.
[31] J.C. Wang, L. Zhang, W.X. Fang, J. Ren, Y.Y. Li, H.C. Yao, J.S. Wang, Z.J. Li, Enhanced

ACCEPTED MANUSCRIPT
photoreduction CO2 activity over direct Z-scheme alpha-Fe2O3/Cu2O heterostructures under
visible light irradiation, ACS Applied Materials & Interfaces, 7 (2015) 8631-8639.
[32] S. Qin, F. Xin, Y. Liu, X. Yin, W. Ma, Photocatalytic reduction of CO2 in methanol to methyl
formate over CuO-TiO2 composite catalysts, Journal of Colloid and Interface Science, 356
(2011) 257-261.

ACCEPTED MANUSCRIPT
Tables
Table 1 The parameters obtained by simulation with the equivalent
circuit model Rs(RctQ)

ACCEPTED MANUSCRIPT
Table 1
Reduction potential /V
Rs /Ω cm2 Rct /Ω cm2
Q/Ω-1·cm-2·s-n
2.19E-5
n
0
8.60
8.72
1.08E4
0.54E4
0.41E4
0.46E4
0.895
0.931
0.925
0.922
-1.2
-1.5
-1.7
2.02E-4
11.18
10.13
2.42E-4
2.60E-4

ACCEPTED MANUSCRIPT
Figures
Fig. 1 CV curve of TNTs in 1mol/L (NH4)2SO4 at scan rate of 50 mV/s.
Fig. 2 EIS plots of various TNTs measured at 0.2 V with 10 mV amplitude sinusoidal
perturbation in 1mol/L CuSO4 solution.
Fig. 3 Ti2p XPS spectra of the TNTs reductive doped at different potentials.
Fig. 4 XRD patterns of TNTs and CuO/reductive doped TNTs electrodes.
Fig. 5 Top and cross-section views of the CuO/TNTs (a, b) without reductive doping
and (c, d) reductive doped at -1.5 V.
Fig. 6 TEM and HRTEM analysis of CuO/TNTs (a,b) with reductive doping and (c)
without reductive doping.
Fig. 7 Schematic illustration of Cu electrodeposition on (a) undoped TNTs and (b)
self-doped TNTs.
Fig. 8 The Linear Sweep Voltammetry of CuO/TNTs electrode in N2-saturated and
CO2-saturated 0.1 M Na2SO4 solution with the scan rate of 50 mV/s under light
on/off.
Fig. 9 Transient photocurrent response of CuO/TNTs at -0.5 V under visible light, (a)
TNTs is reductive doped at different potentials for 60 sec, (b) TNTs is
reductive doped at -1.5 V for different time.
Fig. 10 Methanol yield by using different CuO/TNTs electrodes.
Fig. 11 Methanol and formaldehyde concentrations at different reductive potentials
within 5h.
Fig. 12 Mechanism of photoelectrocatalytic reduction of CO2 on CuO/TNTs.

ACCEPTED MANUSCRIPT
Fig. 1
-0.025
-0.020
-0.015
-0.010
-0.005
0.000
0.005
0.010
Ti(ꢀ)ꢁTi(ꢂ)
0.0 -0.2 -0.4 -0.6 -0.8 -1.0 -1.2 -1.4 -1.6
Potential / V vs. SCE

ACCEPTED MANUSCRIPT
Fig. 2
without reductive doping
reductive doping at -1.2 V
reductive doping at -1.5 V
reductive doping at-1.7 V
-8
-6
-4
-2
0
Rs
Rct
0
2
4
6
8
10
12
Zre / kꢀ cm2

ACCEPTED MANUSCRIPT
Fig. 3
without reductive doping
reductive doping at -1.2 V
Ti 2p
458.5eV
reductive doping at -1.5 V
reductive doping at -1.7 V
458.5eV
458.6eV
458.8eV
466
464
462
460
458
Binding Energy (eV)

ACCEPTED MANUSCRIPT
Fig. 4
★： Ti
◆： CuO
▼: TiO2
CuO/TNTs
TNTs
★
★
★
★
★
★
★
◆
◆
▼
20
40
60
80
2θ （ Degree）

ACCEPTED MANUSCRIPT
Fig. 5

ACCEPTED MANUSCRIPT
Fig .6

ACCEPTED MANUSCRIPT
Fig .7

ACCEPTED MANUSCRIPT
Fig. 8
-5
-4
-3
-2
-1
0
CuO/TNTs
CO light
2
CO dark
2
N
light
2
2
N
dark
-0.59V
0.0
-0.1 -0.2 -0.3 -0.4 -0.5 -0.6 -0.7 -0.8
Potential / V

ACCEPTED MANUSCRIPT
Fig. 9
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
without reductive doping
a
reductive doping at -1.2 V
reductive doping at -1.5 V
reductive doping at -1.7 V
Pure titanium plate
Light off
Light on
0
20
40
Time / s
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
without reductive doping
reductive doping for 5 s
reductive doping for 10 s
reductive doping for 30 s
reductive doping for 60 s
b
Light off
Light on
0
20
40
Time / S

ACCEPTED MANUSCRIPT
Fig. 10
35
30
25
20
15
10
5
No self-doped
Self-doped
0
0
1
2
3
4
5
Time / h

ACCEPTED MANUSCRIPT
Fig. 11
30
25
20
15
10
5
Methanol
Formaldehyde
0
-0.2
-0.3
-0.4
-0.5
-0.6
Potential / V

ACCEPTED MANUSCRIPT
Fig. 12

ACCEPTED MANUSCRIPT
Highlights
ꢀ CuO/TNTs composite electrode with an embedded structure was successfully
fabricated via using self-doped TNTs.
ꢀ The activated surface of TNTs by reductive doping benefits the deposition of Cu
nanoparticles.
ꢀ High activity and stability for CO2 reduction are achieved with the
CuO/self-doped TNTs electrode.
ꢀ The yields of methanol produced by CuO/self-doped TNTs electrode increased
significantly.