Efficient Photoelectrochemical Reduction of CO2 on Pyridyl Covalent Bonded Ruthenium(II) Based-Photosensitizer

Article abstract of DOI:10.1016/j.electacta.2016.08.135

Photo/electrochemical CO₂ reduction using pyridine was feasible to produce methanol via the formation of pyridiniumformate intermediate. To improve the reduction efficiency, a pyridyl bonded ruthenium (II)-based photosensitizer catalyst (Ru-Py) was designed for photoelectrochemical CO₂ conversion. The photocurrent density on Ru-Py modified electrode in CO₂ saturated solution was 0.103?mA?cm^?2 higher than that without illumination. The total Faradaic efficiency (f) reached 83.1%, whereas the turnover number (TON) for methanol was 38.4 in aqueous solution after 8?h irradiation. The methanol production was 24.1?μmol which was higher than the published literatures (less than 8?μmol) in similar systems could be attributed to the efficient electron transfer between the photosensitizer and the pyridyl active site covalently linked by C-C bond, as well as the strong and wide absorption up to 610?nm resulted from the large conjugated structure of the ligands. The mechanism investigation revealed that the N atom in pyridyl as catalytic active sites played significant role in CO₂ conversion by forming the pyridiniumformate intermediate which was confirmed by the simulation reaction. Meanwhile, in order to realize the reduction process intuitively, the density functional theory (DFT) was applied to simulate the structure of Ru-Py and the pyridiniumformate intermediates.

Full text of DOI:10.1016/j.electacta.2016.08.135

Accepted Manuscript
Title: Efficient Photoelectrochemical Reduction of CO2 on
Pyridyl Covalent Bonded Ruthenium (II)
Based-Photosensitizer
Author: Jibo Liu Huijie Shi Xiaofeng Huang Qi Shen Guohua
Zhao
PII:
S0013-4686(16)31870-9
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.electacta.2016.08.135
EA 27913
To appear in:
Electrochimica Acta
Received date:
Revised date:
Accepted date:
27-6-2016
26-8-2016
28-8-2016
Please cite this article as: Jibo Liu, Huijie Shi, Xiaofeng Huang, Qi Shen,
Guohua Zhao, Efficient Photoelectrochemical Reduction of CO2 on Pyridyl
Covalent Bonded Ruthenium (II) Based-Photosensitizer, Electrochimica Acta
http://dx.doi.org/10.1016/j.electacta.2016.08.135
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 Photoelectrochemical Reduction of CO2 on Pyridyl
Covalent Bonded Ruthenium (II) Based-Photosensitizer
Jibo Liu, Huijie Shi, Xiaofeng Huang, Qi Shen, Guohua Zhao*
School of Chemical Science and Engineering, Shanghai Key Lab of Chemical
Assessment and Sustainability, Tongji University, 1239 Siping Road, Shanghai
200092, China
AUTHOR INFORMATION
Corresponding Author
*G. Zhao. Phone: (+86)21-65981180. Fax: (+86)21-65982287.
E-mail address: g.zhao@tongji.edu.cn
1 / 38

Graphical abstract
2 / 38

Highlights




High total faradic efficiency of PEC CO2 conversion was obtained on Ru-Py.
The catalytic activity of pyridyl for CO2 reduction was well retained.
The C-C single bond linkage in Ru-Py facilitates methanol production.
The intermediate pyridiumformate was proved by the simulated reaction.
ABSTRACT:
Photo/electrochemical CO2 reduction using pyridine was feasible to produce methanol
via the formation of pyridiniumformate intermediate. To improve the reduction
efficiency, a pyridyl bonded ruthenium (II)-based photosensitizer catalyst (Ru-Py)
was designed for photoelectrochemical CO2 conversion. The photocurrent density on
Ru-Py modified electrode in CO2 saturated solution was 0.103 mA cm-2 higher than
that without illumination. The total Faradaic efficiency (f) reached 83.1%, whereas
the turnover number (TON) for methanol was 38.4 in aqueous solution after 8 h
irradiation. The methanol production was 24.1 μmol which was higher than the
published literatures (less than 8 μmol) in similar systems could be attributed to the
efficient electron transfer between the photosensitizer and the pyridyl active site
covalently linked by C-C bond, as well as the strong and wide absorption up to 610
nm resulted from the large conjugated structure of the ligands. The mechanism
investigation revealed that the N atom in pyridyl as catalytic active sites played
significant role in CO2 conversion by forming the pyridiniumformate intermediate
which was confirmed by the simulation reaction. Meanwhile, in order to realize the
reduction process intuitively, the density functional theory (DFT) was applied to
3 / 38

simulate the structure of Ru-Py and the pyridiniumformate intermediates.
Keyword: Ru(II) photosensitizer, Pyridyl catalytic site, CO2 reduction,
pyridiniumformate
1. Introduction
Carbon dioxide concentration in the air grew rapidly during the past decades due to
the excess consumption of fuel resource such as coal, petroleum and natural gas and
further caused the resource shortage [1-3]. Efforts has been made to resolve these
problems by converting carbon dioxide to higher-energy species such as formic acid,
methanol or other clean resource [4-12]. Especially, methanol is an important material
for the organic chemical industry and a potential alternative to fossil fuels [13-14].
However, the chemical reduction of CO2 to multi electron reduced liquid product
(CH3OH) remains a challenging task [15-16].
Up to now, metal, metal oxide, metal-organic frameworks (MOFs) and metal
complexes based catalysts have been designed and synthesized for converting CO2 to
desirable products under mild condition [17-26]. Besides, Pyridine is one of the small
organic molecules which is able to catalyze the conversion of CO2 to various C1
compounds effectively through multi-electron/proton transfer processes [27-30].
Thanks to the relatively large electronegativity of N atom in pyridine, CO2 could be
easily bonded to the molecule and thus be activated [31]. Theoretical calculation has
proved that the pyridiniumformate complex (Py-COOH) was a key intermediate for
methanol production [32-34]. In 5-10 mM pyridine aqueous solution, pyridine showed
efficient catalytic capacity to methanol generation on some noble metal electrodes (Pt,
4 / 38

Pd) due to the surface hydrogen on noble metal [15, 33] and limited kinds of
semiconductor electrodes such as n-GaP, p-GaAs and p-InP [29-30, 35]. Pyridine and
its derivatives have also been used in photocatalytic system for CO2 reduction by
combining with metal organic complex as photosensitizer which exhibited excellent
photochemical properties, like suitable redox-properties in both ground and excited
state, strong absorption in the visible light region, photochemical stability, and
sufficiently long-excited state lifetimes [17-18, 36-37]. MacDonnell and co-workers
found that in the catalytic system with Ru(phen)32+ as photosensitizer and pyridine as
catalyst units (Ru(phen)32+:Py = 1:200), the CO2 reduction efficiency was
significantly enhanced, though methanol selectivity was only 0.1% [38]. A
ruthenium(II) polypyridyl sensitizer modified with a pyridyl functional group was
then developed and a better selectivity (9%) to methanol was achieved in non-aqueous
solution because of the more efficient intramolecular electron migration in sensitizer
coupled pyridyl system [39]. Not only the catalytic activity was well kept but also
some unfavorable characteristics caused by dissociative pyridine was apparently
suppressed by covalent chemical bonding pyridyl catalyst to metal complex
photosensitizer. However, the prominent disadvantage was that the reaction with this
catalyst was only performed in organic solution [17, 21-22]. Besides, the formation of
Ru(II) dimer originated from the intermolecular coupling in homogeneous also
decreased the utilization ratio of photogenerated electrons [16]. To avoid the
dimerization and improve the catalytic stability, immobilization of the
photo/electrocatalyst onto electrode surface for photoelectric catalysis application has
5 / 38

also been reported [18, 40].
Hence, in this work, a Ru(II) centered complex Ru(bpy)2(PIP)2+ (bpy =
2-(pyridine-2-yl)pyridine, PIP = 2-(pyridin-3-yl)-1H-imidazo[4,5-f][1,10]phenanthro-
line) (Ru-Py) was designed and immobilized on the fluorine-doped tin oxide coated
glass (FTO) for electrochemical (EC) and photoelectrochemical (PEC) CO2 reduction
in aqueous solution, in which Ru(II) bipyridine complex was employed as
photosensitizer and un-coordinated pyridyl as catalytic units. In order to keep the
independence of the pyridyl catalytic site, a covalent C-C single bonding was
designed to connect Ru(II) photosensitizer and pyridyl catalytic site. PEC CO2
reduction was then performed on Ru-Py modified electrode. To elucidate the
mechanism of CO2 reduction, the influence of pH value, the role of N atom in pyridyl
and the steric hindrance caused by the position of pyridyl-N catalytic sites were all
investigated in detail as controlled experiments. Besides, density functional theory
(DFT) calculation was also applied to optimize the structure of Ru-Py catalyst and
identify the reduction intermediates which was further confirmed by simulation
reaction and the possible mechanism was proposed.
2. Experimental
2.1. Chemicals and Materials
RuCl3.xH2O was purchased from Aladdin Industrial Inc., isonicotinaldehyde (98%)
was purchased from J&K Scientific Ltd. 1,10-phenanthroline (99.8%), H2SO4
(98.3%), HNO3 (68%), 2,2’-bipyridine (99.5%), ammonium acetate (98%), potassium
bromide (99.5%), N,N-dimethylformamide (DMF, 99.5%), potassium chloride
6 / 38

(99.5%), sodium hydroxide (99.8%) and ethanol (99.8%) were purchased from
Sinopharm Chemical Reagent Co., Ltd., SCRC, China without any further dispose.
Note that the ultrapure water with its conductivity reaching 18.2 MΩ cm was used for
all the solutions preparation.
2.2. Synthesis of the Photoelectrocatalyst
The precursor [Ru(bpy)2Cl2 (bpy = 2,2′-bipyridine)] was synthesized by the method
described in the literature [35] from RuCl3.xH2O and 2,2′-bipyridine. The ligand PIP
containing the pyridyl active site was obtained via two steps synthesis from
1,10-phenanthroline followed the method described in the literature [41-42]. The
detail synthesis procedures were described in supporting information (SI).
The photoelectrocatalyst (Ru-Py) was synthesized by the similar method
described in the literature [43] from the precursor and PIP. Ru(bipy)2Cl2 (240 mg, 0.5
mmol) and PIP (150 mg, 0.5 mmol) was refluxed in EtOH/H2O (10:1, 5mL) for 5 h
under the protection of N2. The crude product was obtained after ammonium
hexafluorophosphate (163.0 mg, 1.0 mmol) was added. The solid was purified by
silica-gel chromatography (dichloromethane: methanol = 1:0 to 20:1) to give the title
compound as a red needle solid (120 mg, yield: 32.7%). The procedures for Ru-Ph
1
and Ru-Py(o)(o: ortho-) were similar with that for Ru-Py. The HNMR spectra, XPS
information and detail synthesis procedures were described in supporting information
(Fig. S1-S2, SI).
Ru-Py (4 mg, 0.004 mmol) was added to CH3CN (1mL) and then Nafion
solution (20 μL) was added. The mixture was shaken for 5 mins to get an evenly
7 / 38

solution. 0.5 mL of the solution was dripped slowly to the surface of a FTO with an
area of 1 cm × 2 cm and evaporated naturally. The FTO was then washed with
deionized water and dried in vacuum at 40 oC for 12 hours to obtain a Ru-Py thin film
working electrode for PEC reduction.
2.3. Characterization
1H NMR spectra were obtained on a Bruker AVANCE III 400MHz spectrometer
operating at 400 MHz. The electrochemical measurements were obtained on a
CHI660C electrochemical workstation (CH Instruments, Inc., USA) using a
conventional three-electrode system. The steady-state fluorescence spectra of Ru-Py
were recorded on Hitachi, F-7000 fluorescence spectrophotometer. The absorption
spectra were measured by an ultraviolet-visible spectrophotometer (UV−Vis, Agilent
8453). The methanol was detected with Agilent GC-MS (6890N Network GC system
with a 5973 Network Mass Selective detector) by headspace analysis.
The cyclic voltammetry (CV) measurements in organic phase was recorded in
0.01 mM Ru-Py DMF (containing 1.0 M H2O) solution, 0.1 M [(n-Bu4N)(PF6)] was
dissolved as supporting electrolyte, glassy carbon working electrode, AgNO3/Ag
reference electrode and graphite flake counter electrode. The (CV) measurements in
aqueous solution were obtained with the as-prepared electrode coated with Ru-Py as
the working electrode, Ag/AgCl filled with saturated KCl as the reference electrode
and a graphite flake as the counter electrode to avoid the reduction products
reoxidation [18]. CO2 saturated 0.1 M KCl aqueous solution was employed as the
supporting electrolyte to stabilize the transition states involving CO2 reduction in
8 / 38

transition-metal complexes coordinating a CO2 ligand [36]. All the potentials were
referenced to normal hydrogen electrode (NHE) unless stated otherwise. The
amperometric i-t curve was obtained with an interval of 200 s for light on/off under
the light intensity of 20 mW·cm-2 at a constant potential of -1.0 V (light source:
LA-410UV, Hayashi, Japan).
2.4. PEC Reduction of CO2 and Products Detection
Controlled potential photoelectrolysis of CO2 was conducted in a 100 mL homemade
double-chamber reactor separated with Nafion-117 proton exchange membrane.
Typically, 100 mL 0.1 M KCl solution was used as both the anolyte and catholyte
electrolyte solution. Before PEC reduction, high purity CO2 (99.99%) gas was
bubbled through the KCl solution in the cathode chamber for 40 min at a flow rate of
30 mL min-1 to completely remove dissolved oxygen. Under such condition, CO2
reached saturation in the electrolyte, and the concentration of free CO2 was reached
0.037 M with the pH of 4.1. The potential during the photoelectrochemical reduction
was kept constant at -1.0 V under 100 mW cm−2 irradiation (light source:
PLS-SXE300 xenon lamp, Beijing PerfectLight Co., Ltd., China, with AM 1.5 filter).
The reduction products in aqueous phase were determined by Nash’s reagent
method [18, 44]. Nash reagent was prepared by adding 25.0 g of ammonium acetate,
2.1 mL of acetic acid and 0.2 mL of acetylacetone into water and making the total
volume of the solution 100 mL. Then 2.0 mL of liquid sample was mixed with 2.0 mL
o
of Nash reagent and shaken for 1 h at 60 C. The final solution was analyzed by
UV-vis spectroscopy and the absorbance at 413 nm was used for quantification (The
9 / 38

standard curve was shown in Fig. S2, SI) to give the amount of formaldehyde
(Nformaldehyde). For detection of the amount of formic acid, 0.5 mL of liquid sample was
added into magnesium powder (50 mg) following by drop-wise addition of 0.5 mL
o
37% hydrochloric acid (10 M) at 0 C and then 3 mL 1 M sodium hydroxide. The
resultant suspension was centrifuged at 10 000 rpm for 5 min and 2 mL of supernatant
was mixed with 2 mL of Nash reagent and shaken for 1 h at 60 oC for UV-vis analysis
to determine the total amount of formaldehyde and formic acid (Ntotal). The amount of
formic acid can be determined after subtracting the amount of formaldehyde in the
product (Ntotal - Nformaldehyde). A headspace GC-MS was used to detect the production
of methanol. The chromatographic column (SHRX105MS, 30-m length and 0.5 mm
inner diameter) at an oven temperature of 40 °C was employed in combination with a
MS detector at 230 °C and helium was the carrier gas. Detection at m/z = 31 was
chosen to identify methanol as the reaction products. The samples were preheated at
90 °C in a 20 mL headspace vial with a septa cap, and 1.0 mL of the head space gas
was injected from a syringe heated to 90 °C and analyzed in the GC/MS instrument.
Control samples containing known concentrations of methanol (in the range 40-280
mM) were analyzed to obtain a standard curve (As shown in Fig. S4, SI). To detect
methane, CO and other gaseous product, a thermal couple detector (TCD) and 5 Å
molecular filled column equipped Techcomp GC7900 (Techcomp, China) gas
chromatogram was employed.
3. Results and Discussion
3.1 EC and PEC Characterization of the Ru-Py Catalyst for CO2 Reduction.
10 / 38

The electrochemical response of the Ru(II) photoelectron- catalyst were probed by
CVs in N2 or CO2-saturated organic media (1M H2O in DMF ) containing 0.1 M
tetrabutylammonium hexafluorophosphate (n-Bu4NPF6) as supporting electrolyte and
0.1 M KCl aqueous solution respectively. In N2 saturated organic system, Ru-Py
undergoes sequential, reversible 1e reductions at E1/2 = -0.95 V, -1.14 V and -1.41 V,
respectively (Fig. 1a) which arise from 1e reduction of the two bpy ligands and the
PIP ligand [45-46]. In N2 saturated aqueous solution, the first two waves coalesced to
a single reduction wave with the cathodic peak potential of -1.15 V that arises from
the 2e reduction of the 2 bpy ligands (Fig. 1b). In CO2 saturated solution, the
reduction waves were enhanced observably, with onset potential at about -1.05 V both
in DMF and aqueous solution, indicating the catalytic capability to reduce CO2. To
further decrease the reduction potential, the pH adjustment was involved to optimize
the CO2 reduction potential in aqueous solution. As shown in Fig. 1b, the onset
potential observed for the CO2 reduction shifted positively from -1.05 V to about
-0.79 V accompanied with the pH value changing from 7.0 to 5.0. As in more acidic
aqueous solution (pH<4), the CVs exhibited obvious irreversible feature (Fig. S5, SI).
Upon light irradiation, the peak current density in CO2 saturated aqueous solution at
pH 5.0 was 1.3-folds higher than that without irradiation and the onset potential was
also positively shifted by 0.05 mV under irradiation. These results suggested that the
weakly acidic solution and the introduction of solar light were in favor of CO2
reduction.
The PEC performance of the catalyst Ru-Py was also investigated to illustrate the
11 / 38

photochemical CO2 reduction performance. The cathodic current variation was
monitored by amperometric i-t responses recorded at -1.0 V with chopped light
irradiation with the title compound loaded on FTO as working electrode. As shown in
Fig. 2a, In N2 saturated KCl solution, the photocurrent was also observed due to the
light absorbance of Ru(II) photosensitizer and ligand reduction [17]. The electrons in
the highest occupied molecular orbital (HOMO) energy level was excited and
migrated to the lowest unoccupied molecular orbital (LUMO) energy level due to the
solar harvest to generate cathodic photocurrent under bias potential. In CO2 saturated
KCl solution, the photocurrent response of cathode increased quickly in 80 s light
illuminating duration. This response was attributed to the electrons excitation from
HOMO energy level to LUMO energy level after Ru-Py absorbing the optical energy
[47]. The increment of current density during the light on/off was about 0.103 mA
cm-2 which was much higher than that produced on some other material (as shown in
table 1) under similar conditions [24, 44, 48-50]. The reason can be attributed to the
highly efficient electron migration of the conjugated system. As UV-Vis diffusive
reflectance (UV-DRS) result indicated, a broad absorption bands in the visible range
was observed for the title complex at wavelength of 300-610 nm attributed to the
Ru-N (metal-ligand) transition and π-π* transition (Fig. 2b). Moreover, the increment
of photo-current density in CO2-saturated solution were 2.5 folds higher than that in
N2 saturated solution, suggesting the title compound was provided with the potential
for PEC CO2 reduction.
To investigate the electron transfer route during the CO2 reduction process,
12 / 38

HOMO-LUMO orbital energy level was estimated according to the literatures [51-53].
Using the RuIII/II oxidation potential (Eox vs. NHE), the intersection of the normalized
UV-visible absorption spectrum and the photoluminescence, one can evaluate the
position of the HOMO and the LUMO energy of the Ru(II) complex based catalyst.
The maximum absorption wavelengths of DRS was 40 nm red shift than that UV-vis
absorption spectra. The may attributed to the concentration of the two measuring
method. In DRS, the intermolecular π-π accumulation greatly promoted the electrons
transition and further benefit to absorb long wavelength light [54]. The HOMO orbital
energy equals to the RuIII/II redox potential (Eox vs NHE) approximatively. Eox (1.16 V)
was estimated from CV which was recorded in 0.05 mM Ru-Py DMF solution with
n-Bu4NPF6 (0.1 M) as supporting electrolyte just as shown in Fig. 3a. The LUMO
orbital energy was obtained from the Eox minus the energy gap (E00). E00 (2.26 V) was
calculated based on the equation (E00 = 1240/λ) where the wavelength (λ = 548 nm)
was estimated from the intersection of normalized UV-visible absorption spectrum
and normalized fluorescence spectrum which were detected in 0.05 mM acetonitrile
solution as shown in Fig. 3b. As a result, it could be obtained that the energy levels of
HOMO and LUMO were loaded at 1.359 eV and -0.901 eV. An enough driving force
for the CO2 reduction is achieved because that the reduction potentials for CO2 to
methanol (-0.38 eV) and other products are all positive than the energy levels of
LUMO (As shown in Fig. 3c). Upon light irradiation, the excited photoelectrons from
the HOMO injected into the LUMO through the C-C single bond bridge, and further
migrate to the activated CO2.
13 / 38

3.2. Enhanced PEC CO2 Reduction in Aqueous Solution
PEC reduction of CO2 was conducted in 100 mL of CO2-saturated 0.1 M KCl aqueous
solution under solar light irradiation at −1.0 V. After 8 h irradiation, the liquid
products detected in the final products were formic acid and methanol while methane,
CO and other compounds were not detected without any gas products detected. The
yield of methanol was noticeable though formate acid was the dominant reduction
product in the reaction. The Faradaic efficiency (η) and TON was obtained as the
following equations:
η = (mp×n×F)/(I×t) (1)
TON = mp/mcat
(2)
e
TON (in e) = mp /mcat (3)
where mp is product amount (mol), n is transferred number of electrons, F is Faraday
constant (96485.34 Cmol-1), I is total observed current (A), t is reaction time (s), mp
e
is electrons migration amount (mol) and mcat is catalyst quantity (mol).
As the desired product, the methanol production was 24.1 μmol (Fig. 4a) with a
TON of 3.7 and a Faradaic efficiency of 27.3% under PEC condition. Such a
production rate of methanol was quite higher than most reported system under
photo/electrode condition containing 10 mM pyridine as shown in table 2 [30-32,
55-56]. In such covalent bonding system, the catalytic activity of the pyridyl was well
maintained and the total Faradaic efficiency was 83.1%, which was much higher than
that in pyridine and pyridine derivatives on noble metal electrodes system (ranged
14 / 38

from 14% to 51%) [52]. To evaluate the contribution of PC and EC, three parallel
reactions were introduced respectively (As shown in table 3). In EC system, 9.3 μmol
methanol (TON: 2.4, Faradaic efficiency: 13.2%) was detected while only 2.7 μmol
methanol (TON: 0.7) was obtained under PC condition with ascorbic acid (AA)
present as electron donor. Only a trace of methanol was obtained under PC condition
in the absence of AA, indicating that the plentiful electron supplement was crucial for
the CO2 reduction. The yield of the product under EC condition was higher than that
under PC condition with AA. It was probably due to the EC process that can supply
more electrons than that in PC system whose electron source is sacrificial agent for
the reduction reaction. The electrode not only catalyzed the EC conversion of CO2
directly but also functioned as electron source for PC CO2 conversion under PEC
condition. The synergistic effect of PC and EC in PEC condition caused 2.5 folds
augment for methanol production compared with that in sole PC and EC.
To investigate the crucial active catalytic site of pyridyl in the process of CO2
reduction, Ru-Ph (pyridyl was replaced by benzene) was designed to compare with
Ru-Py. The PEC CO2 reduction results found that only 37.2 μmol formic acid was
detected after 8 h irradiation at -1.0 V and there was not any methanol (the purple line
in Fig. 4b) detected. This indicated that the N atom in pyridyl was essential for the
production of methanol. Besides, the result also verified the N atoms in imidazole
were not the catalytic active site for the CO2 reduction.
In the photoelectrochemical systems, the pH of the electrolyte solution was an
important factor for methanol production. The yield of methanol were 6.2 μmol at pH
15 / 38

= 6.0 (TON: 1.6), 2.9 μmol at pH = 4.0 (TON: 2.9) and 24.1 μmol at pH = 5.0 (TON:
6.4) after 8 h irradiation. As well known, the pKa of pyridinium ion is 5.3 [31]. The
yield of methanol was deeply affected by both the protonated and deprotonated
pyridyl for the overall process. To testify the formation of protonated pyridyl, the
UV-vis spectrophotometer was employed to detect the absorbance spectra variation at
different pH value as shown in Fig. 5a. In the solution of pH = 5.0, the absorbance
between 300-400 nm was intense compared with that at pH = 6.0, which
corresponding to the low-energy absorption of the metal-to-ligand (Ru(dπ)/PIP(pπ*))
charge transition. The absorption change was caused by the protonation of pyridyl
(Ru-Py_H+). The formation of Ru-Py_H+ was in favor of the CO2 insertion to produce
the pyridiniumformate intermediate (The intermediate was detected by the differential
UV-vis absorbance spectra as shown in Fig 5c which would be discussed later) which
gave the final product through the further reduction. On the other hand, the solubility
of CO2 decreased obviously in strong acidic aqueous solution than that in basic
solution. As a result, the optimum methanol production occurring around the pKa of
pyridinium ion (5.3) which was in coincidence with the CVs result.
The electron transfer efficiency between the photo-sensitizer and catalytic site
was also an important factor for the CO2 reduction. As literature reported [38], in the
2+
non-bridged system [Ru(phen)3 photosensitizer with pyridine catalyst], 200-fold
amount of catalyst was needed than Ru(II) photosensitizer to capture the
photo-generated electrons and only 1.7 μmol methanol was detected. As catalyst was
added equally with photosensitizer in the same condition, only 0.17 μmol methanol
16 / 38

was detected. Herein, similar result was also obtained with photosensitizer-catalyst
separated system [Ru(II):Py]. The methanol production was 3.8 μmol (TON = 0.8)
after 8 hours irradiation at -1.0 V potential. In contrast, the yield of methanol in Ru-Py
system were 16 folds than that in [Ru(phen)32+:Py] system and 7.2 folds than that
in[Ru(II):Py] respectively. The highly effective methanol production might be to the
more effective electron migration intramolecular than intermolecular. In the
photosensitizer bonded catalyst system, the photogenerated electrons of
photosensitizer in the excited state transferred to catalytic site (N atom in pyridyl)
through the C-C single bond bridging. The intramolecular electron transfer efficiency
was much higher than intermolecular. The selectivity to methanol was also increased
compared with other system which was also equipped with pyridyl catalytic site. It is
more selective for methanol on both an absolute quantum yield basis and relative
basis rather than the Ru(II) pyrido-pyrazine catalyst [39] (25.7% of the reducing
equivalents end up in methanol for Ru-Py versus 9% for the Ru(II) pyrido-pyrazine
catalyst). These comparisons need to be considered with the notion that solvent
systems are different in the two cases. The higher methanol selectivity in this system
maybe caused by the decreased rigidity of Ru-Py connected by C-C single bond to a
certain extent. On the other hand, the independence of pyridyl was well maintained
than in the pyrido-pyrazine system.
To evaluate the reduced product caused by the different position of the N atom in
pyridyl, Ru-Py(o) was synthesized for CO2 reduction. Headspace GC/MS analysis of
aliquots collected after 8 h photoelectrolysis revealed that only 3.7 methanol per
17 / 38

Ru-Py(o) were produced while 6.4 methanol per Ru-Py were produced. The yield of
methanol for Ru-Py was 1.7 folds than that for Ru-Py(o). The probable reason could
be the unstable pyridiniumformate structure that produced from Ru-Py(o) and
activated CO2. The pyridiniumformate internidate formed from Ru-Py and activated
CO2 was consummate co-planar surface and all the atoms were in the same plane (Fig
6a) while the intermniadate structure of Ru-Py(o) was unduly tortile (Fig 6b). The
DFT optimized pyridiniumformate showed the dihedral angle that built by the CO2
loaded surface and pyridyl loaded was 18o (angle 2 in Fig. S6) while the dihedral
angle that built by the pyridyl loaded surface and phenanthroline loaded was 163o
(angle 1 in Fig. S6). The tension of the tortile intermediate structure caused the
instability of the pyridiniumformate and further resulted the decomposition of the
intermediate to obtain the 2e product formic acid without deeply reduction to give 6e
reduced product.
3.3 Intermediate Determination and the Plausible CO2 Catalytic Reduction Pathway
DFT calculations were carried out by using the Gaussian 09 package of programs to
further elaborate the feasibility for forming pyridiniumformate as intermediate.
B3LYP functional was used with LanL2DZ basis set for Ru atom and 6-31G++(d, p)
for H, C, N, O atoms [57-61].
The optimized structure of Ru-Py showed that the coordination mode of the Ru
atom is regular octahedron with 4 N atoms from two 2,2’- bipyridine and 2 N atoms
from ligand PIP. The Ru atom was located at the center of the octahedron while the 6
N atoms was dispersed on the apexes. The length of Ru-N bonds range among 2.102
18 / 38

Å - 2.119 Å. The PIP ligand was fully planar with all the atoms in the same plane to
keep the conjugation of the π electrons that in favor of the electron migration from Ru
photosensitizer unit to the pyridyl active site. The HOMO of Ru-Py was populated on
the uncoordinated pyridyl ring of PIP ligand whereas the LUMO resided primarily on
the π system of the bpy moieties (Fig. 6c-d). The carbon atom in CO2 is
electropositive and it is beneficial for the CO2 getting close to the pyridyl catalytic site
of the Ru(II) complex to form the pyridiniumformate [61]. After Ru-Py was
protonated in acidic environment (Ru-Py_H+), the main distribution of the HOMO is
the two bpy and the PIP ligand, meanwhile the LUMO of Ru-Py_H+ were mainly
distributed on the bpy ligands and the uncoordinated pyridinium ring. The electron
density distribution may facilitate electron injecting to the activated CO2 molecule
because it provides electronic hybridizing between the photosensitizer and the
pyridinium catalytic unit. The calculated LUMO level (-0.72 eV in vacuum, 4.8 eV
theoretical absolute potential was used) was also matched with experimental value
(-0.91 eV in DMF/H2O). Compared with -0.38 eV (CO2/MeOH theoretical reduction
potential), it further demonstrated the electron preferred injection from the
photosensitizer to activated CO2 by way of pyridinium catalytic unit.
To futher illustrate the probable mechanism for CO2 reduction, the differential
UV-vis absorbance spectra (ΔA spectra) was employed to detect the possible
intermediates pyridiniumformate in 0.1 M KCl aqueous solution as shown in Fig. 5d.
The absorbance of the CO2 saturated aqueous solution was selected as baseline,
whose pH value was adjusted to 5.0 by 1 M HCl aqueous solution. Downward peaks
19 / 38

indicated bands disappearing while upward counterparts corresponded to new bands
appearing due to photolysis. The change of the absorbance at 364 nm and 520 nm
were associated to the formation of key intermediate pyridiniumformate and the dimer
(Ru-Py)2 respectively. The change of the absorbance between 400 - 450 nm attributed
to the decomposition of the catalyst [17, 38] and the weakening of M-L bonds by the
formation of pyridiniumformate.
The new absorbance bands at 364 nm was caused by the formation of
pyridiniumformate intermediate which was the key factor for the production of
methanol [39]. To verify the formation pyridiniumformate during the reduction
process, a simulate reaction was conducted as follows. Methyl chloroformate which
does not display absorbance at 300 - 800 nm in acetonitrile solution (the dash line in
Fig. 5b) was employed to simulate the pyridiniumformate because it can easily react
with the exposed N atom in pyridyl to give the pyridiniumformate spontaneously
without any further processing [62]. The absorbance change of the UV-vis spectra at
340 nm - 400 nm after the addition of the methyl chloroformate can be attributed to
the generation of pyridiniumformate (the blue line and the purple line in the Fig. 5b).
In contrast, the UV-vis spectra of Ru-Ph had no obvious change after the addition of
methyl chloroformate, which indicated the two N atoms in imidazole was difficult to
form the formate complex (Fig. 5c). The change of the UV-vis absorption at 340 nm -
400 nm was only caused by the pyridiniumformate that formed by the C-N bond
which was composed by the C atom from CO2 and N atom from pyridyl. The
time-dependent change of the ΔA spectra showed the pyridiniumformate increased
20 / 38

continually in the first hour and level off in the next irradiation time (as shown in Fig.
5d). This phenomenon was attributed to the balance of pyridiniumformate formation
and dissociation after irradiation for 1 hour. The time-dependent reduced product
amount was consistent with the ΔA spectra within the same time scale
The absorbance bands at 520 nm attributable to photochemical ligand
substitution of the Ru(II) photosensitizer unit giving the corresponding Ru(II)
bisdiimine-type complex. This should be a main reason why the rate of photocatalytic
production using Ru-Py gradually decreased. Radical dimerization of the Ru(II)
photosensitizer unit was reported as the major secondary reaction to quench the
photoinduced Ru(II) radical [17]. The upward peaks in Fig. 5e at 520 nm were
originated from photochemical ligand substitution of the Ru(II) photosensitizer unit
giving that the corresponding Ru(II) bisdiimine-type complex [39]. The dimerization
can be avoided effectively by immobilizing the Ru(II) complexes on the surface of
FTO electrode during CO2 reduction. Fig. 5e showed the ΔA spectra of the Ru(II)
photoelectrocatalyst immobilized on the surface of FTO electrode after 10 mins, 1
hour and 8 hours irradiation. The ΔA spectra was obtained from comparing the Ru(II)
photoelectrocatalyst solutions through soaking the working electrode by ethanol after
different irradiation time. Comparing the ΔA spectra of immobilized Ru(II)
photoelectrocatalyst on the surface of FTO electrode with homogeneous reaction, the
disappearance of absorption band at 520 nm indicated that the dimerization was
avoided effectively in heterogeneous reaction after the irradiation. Therefore,
convenient recycling was another advantage to load the Ru(II) photoelectrocatalyst on
21 / 38

the surface of FTO electrode than diluted in the solution. The absorption bands
between 400 - 450 nm correspond to metal-to-ligand charge transfer transition bands
(MLCT). As the intermediate formed, the electrons in the Ru(II) photosensitizer
migrate to the pyridiniumformate for CO2 reduction thus caused the electron density
reduced. The reduced electron density decreased the metal-to-ligand charge transfer
which was reflected in the UV-vis absorption [37]. Just as shown in Fig. 5e, the
absorption bands between 400 -450 nm was also observed as the Ru-Py loaded on
FTO.
Based on the above information, the mechanism of PEC CO2 reduction was
proposed in Scheme 1. The first step was the spontaneous protonation of the
photoelectrocatalyst to give the protonated compound (Ru-Py_H+) in the weakly
acidic solution (pH = 5.0). Subsequently the photo-induced electron of Ru(II)
photosensitizer transferred to pyridyl catalytic site through the C-C single bond to
obtain the radicals of the catalyst (Ru-Py_H0). The key intermediate (Ru-Py_H0) was
formed following the activated CO2 insertion into the N-H bond of Ru-Py_H0, which
was detected by UV-vis absorbance. Followed by a series of proton-coupled electron
transfer and the release of small molecule, the reduction product such as formic acid,
formaldehyde and methanol were generated. Simultaneously, the photoelectrocatalyst
was regenerated with the generation of the reduction product.
4. Conclusion
Herein, we have designed a covalent C-C bond Ru-Py to convert CO2 to methanol
22 / 38

under PEC system, in which the uncoordinated pyridyl functioned as active site.
Different from most of the previous reports in which Ru(II) complex produced 2e
transfer product (CO, formic acid), the Ru(II) photosensitizer bonded catalytic active
site in this work can convert CO2 to multi-electron reduction products. Uncoordinated
pyridyl as the catalytic site in the system showed powerful ability to convert CO2
yielding more valuable products such as methanol due to the rapid intramolecular
electron transfer. ΔA spectra analysis revealed that the probable reaction route was the
formation of pyridinium-formate intermediate from protonated pyridyl complex
which was influenced by the pH value. DFT verified that the electron density of
HOMO orbital was contributed from the main-ligand containing the uncoordinated
pyridyl that in favor of the formation of pyridinium-formate. The yield of the
reduction product was influenced by the steric hindrance of the pyridinium-formate
intermediate. Although the dimerization of the Ru(II) photoelectrocatalyst was
decreased to a limited extent by immobilizing onto the surface of the FTO electrode,
the major challenge and the problem urgently to be resolved was the optical instability
of the Ru(II) complexes, which limited the extensive using of molecular catalysts.
ASSOCIATED CONTENT
Supporting Information: The detailed information about the ligand synthesis, the 1H
NMR Fig. for Ru-Py, Ru-Py(o) and Ru-Ph, the standard curves for formic acid and
methanol, and Ru 3d3/2 XPS spectra. This material is available free of charge via the
Internet at http://dx.doi.org/10.1016/j.electacta.
Notes
23 / 38

The authors declare no competing financial interest.
ACKNOWLEDGMENTS
This work was financially supported by the National Natural Science Foundation of
China (NSFC, No. 21477085, 21537003) and Science &Technology Commission of
Shanghai Municipality (14DZ2261100).
REFERENCES
1. K. Sekizawa, K. Maeda, K. Domen, K. Koike, O. Ishitani, Artificial Z-Scheme
Constructed with a Supramolecular Metal Complex and Semiconductor for the
Photocatalytic Reduction of CO2. J. Am. Chem. Soc. 135 (2013) 4596-4599.
2. E. E. Benson, C. P. Kubiak, A. J. Sathrum, J. M. Smieja, Electrocatalytic and
Homogeneous Approaches to Conversion of CO2 to Liquid Fuels. Chem. Soc. Rev. 38
(2009) 89-99.
3. M. Aresta, A. Dibenedetto, A. Angelini, Catalysis for the Valorization of Exhaust
Carbon: From CO2 to Chemicals, Materials, and Fuels. Technological Use of CO2.
Chem. Rev. 114 (2014) 1709-1742.
4. H. G. Baldovi, S. Neatu, A. Khan, A. M. Asiri, S. A. Kosa, H. Garcia,
Understanding the Origin of the Photocatalytic CO2 reduction by Au- and Cu-Loaded
TiO2: A Microsecond Transient Absorption Spectroscopy Study. J. Phys. Chem. C 119
(2015) 6819-6827.
5. C. Kim, H. S. Jeon, T. Eom, M. S. Jee, H. Kim, C. M. Friend, B. K. Min, Y. J.
Hwang, Achieving Selective and Efficient Electrocatalytic Activity for CO2 Reduction
Using Immobilized Silver Nanoparticles. J. Am. Chem. Soc. 137 (2015) 13844–13850
24 / 38

6. S. Rasul, D. H. Anjum, A. Jedidi, Y. Minenkov, L. Cavallo, K. Takanabe, A
Highly Selective Copper-Indium Bimetallic Electrocatalyst for the Electrochemical
Reduction of Aqueous CO2 to CO. Angew. Chem., Int. Ed. 54 (2015) 2146-2150.
7. P. Kumar, S. Kumar, S. Cordier, S. Paofai, R. Boukherroub, S. L Jain,
Photoreduction of CO2 to Methanol with Hexanuclear Molybdenum [Mo6Br14]2-
Cluster Units under Visible Light Irradiation. Rsc. Adv., 4 (2014) 10420-10423.
8. H. Q. Xu, J. Hu, D. Wang, Z. Li, Q. Zhang, Y. Luo, S. H. Yu, H. L. Jiang,
Visible-Light Photoreduction of CO2 in a Metal-Organic Framework: Boosting
Electron-Hole Separation Via Electron Trap States. J. Am. Chem. Soc. 137 (2015)
13440-13443.
9. L. Z. Chen, D. D. Huang, J. Z. Ge, F. M. Wang, A Novel Ag(I) Coordination
Polymers Based on 2-(Pyridin-4-Yl)-1h-Imidazole-4,5-Dicarboxylic Acid: Syntheses,
Structures, Ferroelectric, Dielectric and Optical Properties. Inorg. Chim. Acta. 406
(2013) 95-99.
10. T. Funaki, N. Onozawa-Komatsuzaki, K. Kasuga, K. Sayama, H. Sugihara, New
Class of Thiocyanate-Free Cyclometalated Ruthenium(II) Complexes Having a
Pyridylquinoline Derivative for near-Infrared Sensitization of Dye-Sensitized Solar
Cells. Inorg. Chem. Commun. 35 (2013) 281-283.
11. F. Koschany, D. Schlereth, O. Hinrichsen, On the Kinetics of the Methanation of
Carbon Dioxide on Coprecipitated NiAl(O)X. Appl Catal B-Environ 181 (2016)
504-516.
12. D. B. Kayan, F. Koleli, Simultaneous Electrocatalytic Reduction of Dinitrogen
25 / 38

and Carbon Dioxide on Conducting Polymer Electrodes. Appl Catal B-Environ 181
(2016) 88-93.
13. P. Kumar, C. Joshi, N. Labhsetwar, R. Boukherroub, S. L. Jain, A novel Ru/TiO2
hybrid nanocomposite catalyzed photoreduction of CO2 to methanol under visible
light. Nanoscale 7 (2015) 15258 - 15267.
14. J. Xiao, D. S. Mao, X. M. Guo, J. Yu, Methanol Synthesis from CO2
Hydrogenation over CuO-ZnO-TiO2 Catalysts: The Influence of TiO2 Content,
Energy Technol. 3 (2015) 32 - 39.
15. C. H. Lim, A. M. Holder, J. T. Hynes, C. B. Musgrave, Reduction of CO2 to
Methanol Catalyzed by a Biomimetic Organo-Hydride Produced from Pyridine. J. Am.
Chem. Soc. 136 (2014) 16081-16095.
16. Y. Yan, E. L. Zeitler, J. Gu, Y. Hu, A. B. Bocarsly, Electrochemistry of Aqueous
Pyridinium: Exploration of a Key Aspect of Electrocatalytic Reduction of CO2 to
Methanol. J. Am. Chem. Soc. 135 (2013) 14020-14023.
17. A. Nakada, K. Koike, T. Nakashima, T. Morimoto, O. Ishitani, Photocatalytic
CO2 Reduction to Formic Acid Using a Ru(II)-Re(I) Supramolecular Complex in an
Aqueous Solution. Inorg. Chem. 54 (2015) 1800-1807.
18. X. Huang, G. Zhao, J. Liu, N. Yang, CO2 adsorption-enhanced
semiconductor/metal-complex hybrid photoelectrocatalytic interface for efficient
formate production, Energy
10.1039/C6EE00968A.
&
Environmental Science, (2016) DOI:
19. S. Q. Zhang, L. N. Li, S. G. Zhao, Z. H. Sun, J. H. Luo, Construction of
26 / 38

Interpenetrated Ruthenium Metal-Organic Frameworks as Stable Photocatalysts for
CO2 Reduction. Inorg. Chem. 54 (2015) 8375-8379.
20. S. Wang, J. Lin, X. Wang, Semiconductor-Redox Catalysis Promoted by
Metal-Organic Frameworks for CO2 Reduction. Phys. Chem. Chem. Phys. 16 (2014)
14656-14660.
21. S. Zhang, P. Kang, S. Ubnoske, M. K. Brennaman, N. Song, R. L. House, J. T.
Glass, T. J. Meyer, Polyethylenimine-Enhanced Electrocatalytic Reduction of CO2 to
Formate at Nitrogen-Doped Carbon Nanomaterials. J. Am. Chem. Soc. 136 (2014)
7845-7848.
22. S. Zhang, P. Kang, T. J. Meyer, Nanostructured Tin Catalysts for Selective
Electrochemical Reduction of Carbon Dioxide to Formate. J. Am. Chem. Soc. 136
(2014) 1734-1737.
23. A. Nayak, R. R. Knauf, K. Hanson, L. Alibabaei, J. J. Concepcion, D. L. Ashford,
J. L. Dempsey, T. J. Meyer, Synthesis and Photophysical Characterization of
Porphyrin and Porphyrin-Ru(II) Polypyridyl Chromophore-Catalyst Assemblies on
Mesoporous Metal Oxides. Chem. Sci. 5 (2014) 3115-3119.
24. S. Wang, X. Wang, Photocatalytic CO2 Reduction by CdS Promoted with a
Zeolitic Imidazolate Framework. Appl Catal B-Environ. 162 (2015) 494-500.
25. P. Li,; Y. Zhou, Z. Zhao, Q. Xu, X. Wang, M. Xiao, Z. Zou, Hexahedron
Prism-Anchored Octahedronal CeO2: Crystal Facet-Based Homojunction Promoting
Efficient Solar Fuel Synthesis. J. Am. Chem. Soc. 137 (2015) 9547-50.
26. D. Gao, H. Zhou, J. Wang, S. Miao, F. Yang, G. Wang, J. Wang, X. Bao,
27 / 38

Size-Dependent Electrocatalytic Reduction of CO2 over Pd Nanoparticles. J. Am.
Chem. Soc. 137 (2015) 4288-4291.
27. S. Daniel, C. L David, C. P Jonas, Clifford P. K; Reduction of CO2 by Pyridine
Monoimine Molybdenum Carbonyl Complexes: Cooperative Metal–Ligand Binding
of CO2. Chem. Eur. J. 21 (2015) 8497-8503.
28. E. Lebegue, J. Agullo, M. Morin, D. Belanger, The Role of Surface Hydrogen
Atoms in the Electrochemical Reduction of Pyridine and CO2 in Aqueous Electrolyte.
Chemelectrochem 1 (2014) 1013-1017.
29. J. A. Keith, E. A. Carter, Theoretical Insights into Pyridinium-Based
Photoelectrocatalytic Reduction of CO2. J. Am. Chem. Soc. 134 (2012) 7580-7583.
30. E. E. Barton, D. M. Rampulla, A. B. Bocarsly, Selective Solar-Driven Reduction
of CO2 to Methanol Using a Catalyzed P-Gap Based Photoelectrochemical Cell. J.
Am. Chem. Soc. 130 (2008) 6342-6344.
31. A. J. Lucio, S. K. Shaw, Pyridine and Pyridinium Electrochemistry on
Polycrystalline Gold Electrodes and Implications for CO2 Reduction. J Phys Chem C
119 (2015) 12523-12530.
32. J. H. Jeon, P. M. Mareeswaran, C. H. Choi, S. I. Woo, Synergism between Cdte
Semiconductor and Pyridine - Photoenhanced Electrocatalysis for CO2 Reduction to
Formic Acid. Rsc Adv 4 (2014) 3016-3019.
33. A. B. Bocarsly, Q. D. Gibson, A. J. Morris, R. P. L'Esperance, Z. M. Detweiler, P.
S. Lakkaraju, E. L. Zeitler, T. W. Shaw, Comparative Study of Imidazole and Pyridine
Catalyzed Reduction of Carbon Dioxide at Illuminated Iron Pyrite Electrodes. Acs
28 / 38

Catal. 2 (2012) 1684-1692.
34. R. Ma, A. H. Liu, C. B. Huang, X. D. Li, L. N. He, Reduction of Sulfoxides and
Pyridine-N-Oxides over Iron Powder with Water as Hydrogen Source Promoted by
Carbon Dioxide. Green. Chem. 15 (2013) 1274-1279.
35. M. Lessio, E. A. Carter, What Is the Role of Pyridinium in Pyridine-Catalyzed
CO2 Reduction on P-Gap Photocathodes? J. Am. Chem. Soc. 137 (2015)
13248−13251.
36. D. J. Boston, C. Xu, D. W. Armstrong, F. M. MacDonnell, Photochemical
Reduction of Carbon Dioxide to Methanol and Formate in a Homogeneous System
with Pyridinium Catalysts. J. Am. Chem. Soc. 135 (2013) 16252-16255.
3
37. Y. J. Sun, Y. Liu, C. Turro, Ultrafast Dynamics of the Low-Lying MLCT States
of [Ru(bpy)2(dppp2)]2+. Chem. Commun. J. Am. Chem. Soc. 132 (2010) 5594-5595.
38. R. Kuriki, K. Sekizawa, O. Ishitani, K. Maeda, Visible-Light-Driven CO2
Reduction with Carbon Nitride: Enhancing the Activity of Ruthenium Catalysts.
Angew. Chem., Int. Ed. 54 (2015) 2406-2409.
39. D. J. Boston, Y. M. Pachon, R. O. Lezna, N. R. de Tacconi, F. M. MacDonnell,
Electrocatalytic and Photocatalytic Conversion of CO2 to Methanol Using Ruthenium
Complexes with Internal Pyridyl Cocatalysts. Inorg. Chem. 53 (2014) 6544-6553.
40. P. Kang, S. Zhang, T. J. Meyer, M. Brookhart, Rapid Selective Electrocatalytic
Reduction of Carbon Dioxide to Formate by an Iridium Pincer Catalyst Immobilized
on Carbon Nanotube Electrodes. Angew. Chem., Int. Ed. 53 (2014) 8709-8713.
41. X. Wang, J. Zhang, G. Liu, H. Lin, A Series of Cd(II) Complexes with Π–Π
29 / 38