Lewis acid activated CO2 reduction over a Ni modified Ni-Ge hydroxide driven by visible-infrared light

Article abstract of DOI:10.1039/c8dt04408b

Improvement of light harvesting and reaction kinetics is of great importance for achieving efficient solar-driven CO₂ reduction. Here, a Ni modified low-crystalline Ni-Ge containing hydroxide with Lewis acid sites was synthesized in highly redu

Full text of DOI:10.1039/c8dt04408b

Dalton
Transactions
View Article Online
View Journal
PAPER
Lewis acid activated CO reduction over a Ni
2
modified Ni–Ge hydroxide driven by visible-
Cite this: DOI: 10.1039/c8dt04408b
infrared light†
a
a
b
a
c
a
a
Zhenyu Xin,‡ Lei Lu,‡ Bing Wang, Xiaohui Wang, Kai Zhu,* Zhe Xu, Zhentao Yu,
a
a,b
Shicheng Yan * and Zhigang Zou
Improvement of light harvesting and reaction kinetics is of great importance for achieving efficient solar-
driven CO
2
reduction. Here, a Ni modified low-crystalline Ni–Ge containing hydroxide with Lewis acid
solution and exhibited 9.3 μmol gcat.− h−1 CO and
1
sites was synthesized in highly reductive NaBH
4
.5 μmol gcat. h−1 CH
−
1
generation rates under visible light irradiation, and even achieved a 3.8 μmol
CO evolution under infrared light irradiation. The wide-spectrum light harvesting resulted from
3
g
4
−
1
−1
cat.
h
Received 6th November 2018,
Accepted 4th December 2018
the light absorption from the localized surface plasmonic resonance of Ni nanoparticles. In addition, the
Lewis acid can activate CvO bonds to decrease the kinetic barriers of CO reduction. The design
2
DOI: 10.1039/c8dt04408b
concept that rationally combines the advantages of expanding the spectral response and activating CO
2
rsc.li/dalton
may offer a new strategy for efficient solar energy utilization.
tants. Particularly, it has been well demonstrated that the CO2
molecule can be activated by Lewis acids such as coordinately
Introduction
2
Photocatalytic CO conversion into renewable solar fuels is a unsaturated metal ions, generally resulting from the fact that
promising technique for solving energy shortage and global the lone pair electrons of each oxygen atom in CO can be
2
1
7
3+
warming problems. Generally, oxidation of H O into oxygen donated to the surface of Lewis acid centers. Ti as a Lewis
2
and reduction of CO
2
into carbon-based products (CO, CH
4
2
acid on the anatase TiO nanotubes allowed the creation of
8
etc.), respectively, driven by photogenerated holes and elec- bent CO structures, thus benefiting the dissociation of CO .
trons, are involved in this process. Both the half-reactions Similarly, in our previous work, Ge on the α-Zn–Ge–O surface
require high reaction energy and multi-electron consumption, could effectively react with CO to form carbon intermediate
2
2
2
+
2
2
9
thus resulting in slow reaction kinetics. In addition to the species (CO*), and even split CO into C and O . In addition,
2
2
kinetic limitations, the light harvesting of a photocatalyst is a the oxygen vacancies (O ) can serve as an electron mediator to
Vs
1
0–13
thermodynamic limit for the maximum solar-to-chemical facilitate the activation of molecules.
Accordingly, simul-
energy conversion efficiency. These facts indicate that decreas- taneously activating CO by a solid base and H O by O affords
2
2
Vs
14
ing kinetic barriers and increasing light harvesting are highly a two-fold enhancement in CO reduction. This evidence indi-
2
3
–6
necessary for achieving an efficient CO
2
conversion.
2 2
cates that the activation of CO and H O is indeed important
To decrease the kinetic barriers of the CO reduction reac- for CO2 conversion. However, there is a lack of methods to
2
tion, one of the most effective methods is to activate the reac- combine molecule activation with efficient light harvesting.
Herein, we aim to develop a strategy that simultaneously
decreases the kinetic barriers of the reaction and expands the
a
Jiangsu Key Laboratory of Artificial Functional Materials, Eco-materials and
Renewable Energy Research Center (ERERC), National Laboratory of Solid State
Microstructures, Collaborative Innovation Center of Advanced Microstructures,
College of Engineering and Applied Sciences, Nanjing University, 210093,
P. R. China. E-mail: yscfei@nju.edu.cn
light harvesting. To confirm this purpose, a composite photo-
catalyst, a Ni modified low-crystalline Ni–Ge hydroxide with
Lewis acid sites, was synthesized by an ion exchange reaction
between Ni(NO ) and Na GeO in sodium borohydride solu-
3 2
2
3
b
Jiangsu Province Key Laboratory for Nanotechnology, School of Physics,
tion at 273 K. The coordinately unsaturated metal sites (Lewis
Nanjing University, Nanjing, Jiangsu 210093, P. R. China
acid sites) served as active sites for activating CO , and the
localized surface plasmonic resonance (LSPR) effect of the Ni
nanoparticles was used to broaden the optical absorption. A
2
c
School of Information Science and Engineering, Nanjing University Jinling College,
No. 8 Xuefu Road, Nanjing, Jiangsu 210089, P. R. China. E-mail: 25898099@qq.com
†
Electronic supplementary information (ESI) available. See DOI: 10.1039/
c8dt04408b
These authors contributed equally to this work.
remarkable photocatalytic performance in CO reduction was
2
‡
obtained under irradiation by visible and infrared light. We
This journal is © The Royal Society of Chemistry 2018
Dalton Trans.

View Article Online
Paper
Dalton Transactions
have confirmed that the surface –OH as a solid proton on the tion–desorption technique using an automatic surface area
low-crystalline Ni–Ge containing hydroxide could be easily oxi- analyzer (Micromeritics Tristar-3000, USA) after dehydration at
+
dized by photogenerated holes to form H and O
2
and recov- 423 K for 3 h under N
2
flow. CO
2
adsorption was detected at
δ+
2
ered by H O, while the reduced Ge (0 < δ < 4) as a Lewis acid 273 K based on the Brunauer–Emmett–Teller (BET) method at
is responsible for the activation of CO , significantly decreas- P/P0 = 0.03. Mott–Schottky plots were determined using an
2
ing the reaction barriers. Our findings offer a simple route to electrochemical workstation (CHI Instruments CHI760E) at fre-
introduce Lewis acids as active sites and the LSPR effect for quencies of 200, 500, and 1000 Hz in the dark.
effective utilization of visible-infrared light.
Photocatalytic CO
2
reduction tests
The photocatalytic CO
2
reduction tests were conducted on a
glass reactor by dispersing a photocatalyst (50 mg) onto a
sample holder (area, 4.2 cm ). The volume of the reaction
system was about 230 mL. The light source was a 300 W Xe
lamp and the radiation spectrum is shown in Fig. S1.† Before
Experimental section
Synthesis of a Ni modified Ni–Ge containing hydroxide
2
photocatalyst
2 3
To start with, Na GeO powders were prepared by heating a the irradiation, the reaction system was evacuated and flushed
stoichiometric mixture of GeO and Na CO at 1173 K for 12 h. with CO several times, and then high-purity CO (99.999%) was
2
2
3
2
2
After that, NaBH
obtained Na GeO
4
powders (14 mM) were dissolved in the as- introduced into the reaction chamber to achieve an ambient
aqueous solution (2 mM, 10 mL). The pressure. Subsequently, 0.4 mL of deionized water was injected
2
3
resulting mixed solution was then injected into Ni(NO3)2 into the chamber as a reductant, thus achieving a saturated
aqueous solution (3 mM, 20 mL) with ice-water soaking. After vapor pressure (molar gas ratio of CO /H O was about 29 : 1).
2
2
magnetic stirring for 20 h in an ice bath, the resulting powders Prior to irradiation, the as-prepared photocatalyst was main-
were collected by centrifugation, washing several times with tained in the dark for 12 h to reach the adsorption–desorption
deionized water, and freeze-drying for 48 h (denoted as Ni/ equilibrium of CO . The amount of O from photocorrosion or
2
2
Ni
3
Ge
2
O
5
(OH)
4
). For comparison, the same procedure without CO2 reduction was online determined using gas chromato-
the addition of NaBH4 was applied to synthesize a single- graphy (GC-8A, MS-5A column, thermal conductivity detector,
phase Ni
3
Ge
2
O
5
(OH)
4
.
Ar carrier, Shimadzu, Japan). During the reaction, 1 mL gas was
extracted by using a sampling needle from the chamber at given
Photocatalyst characterization
intervals for subsequent CO and CH concentration analysis by
4
gas chromatography (GC-2014, Shimadzu Corp., Japan).
The crystallographic structure was identified by powder X-ray
diffraction (XRD, Rigaku Ultima III, Japan, Cu Kα radiation)
operating at 40 kV and 40 mA. The surface morphology and
composition were analyzed using scanning electron Results and discussion
microscopy (SEM, FEI Nova Nano SEM 230, USA). High-resolu-
tion images and selected area electron diffraction (SAED) pat-
terns were obtained by transmission electron microscopy
X-ray powder diffraction (XRD) analysis (Fig. 1a) showed that a
mixture of Ni (JCPDS no. 04-0850) and Ni Ge O (OH) (JCPDS
3
2
5
4
no. 11-0097) was obtained by a reaction of Ni(NO
3 2
) and
(TEM, FEI Tecnai G2 F30 S-Twin, USA) operating at 200 kV.
Na GeO in NaBH4 solution at 273 K (denoted as Ni/
2
3
Ultraviolet–visible (UV-vis) diffuse reflectance spectra (DRS)
were recorded with a UV-vis spectrophotometer (UV-LAMBDA
Ni Ge O (OH) ). Single-phase Ni Ge O (OH) crystals were
3
2
5
4
3
2
5
4
obtained via a similar synthetic procedure without adding
9
50, PerkinElmer, USA) and were transformed into the absorp-
tion spectra according to the Kubelka–Munk relationship.
Quantitative Fourier transform infrared spectroscopy (FT-IR)
was performed using a Nicolet Nexus 870 infrared spectro-
meter (Nicolet, USA) under ambient conditions. Room-temp-
erature steady-state photoluminescence (PL) spectroscopy and
PL decay traces were obtained by using a fluorescence spectro-
fluorometer (HORIBA Fluorolog-3, HORIBA). The chemical
state and valence band spectra of photocatalysts were investi-
gated by X-ray photoelectron spectroscopy (XPS) on a PHI5000
Versa Probe (ULVAC-PHI, Japan) with monochromatized Al Kα
X-ray radiation (1486.6 eV). The binding energy was deter-
mined by reference to the C 1s line at 284.6 eV. The elemental
Ix=Sx
Fig. 1 (a) XRD patterns of Ni/Ni
The SEM image and (d) TEM image of Ni/Ni Ge O (OH) . The inset
3 2 5 4 3 2 5 4
Ge O (OH) and Ni Ge O (OH) . (b)
content (C ) was determined by a formula Cx ¼ P
, where
x
Ii=Si
3
2
5
4
i
showed the SAED patterns of Ni/Ni
e) TEM image of Ni Ge (OH) . The inset shows the SAED patterns of
Ge (OH)
3 2 5 4
Ge O (OH) . (c) The SEM image and
I is the XPS peak area and S is the sensitivity factor. The
(
3
2
O
5
4
specific surface area was measured by a nitrogen (N ) adsorp- Ni
3
2
O
5
4
.
2
Dalton Trans.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Dalton Transactions
Paper
NaBH
sample prepared by directly mixing the NaBH4 solution and
Ni Ge (OH) (Fig. S2†). However, the Ni nanoparticles could
be obtained by NaBH reduction of Ni(NO . This evidence
suggested that the formation of Ni in Ni/Ni Ge O (OH) had
4
(Fig. 1a). No XRD peaks of Ni could be detected in a
3
2
O
5
4
4
3 2
)
3
2
5
4
the same crystallization process as the reduction of Ni(NO
by NaBH . Energy dispersive spectroscopy (EDS) revealed that
the atomic ratio of Ni : Ge in Ni/Ni Ge O (OH) and
3 2
)
4
3
2
5
4
Ni
The higher Ni content in Ni/Ni
mation of Ni nanoparticles. Compared with Ni Ge O (OH) ,
3
Ge
2
O
5
(OH)
4
was about 1.76 and 1.59, respectively (Fig. S3†).
3
Ge (OH) indicated the for-
2
O
5
4
3
2
5
4
the broadening of XRD peaks indicated a lower crystallinity of
Ni/Ni Ge (OH) . To further confirm that the low-crystallinity
mixture was indeed composed of Ni and Ni Ge O (OH) , the
3
2
O
5
4
3
2
5
4
as-prepared sample was heated at 673 K for 1 h under N
The high-resolution transmission electron microscopy
HR-TEM) image of Ni/Ni Ge O (OH) after annealing treat-
ment showed two lattice spacings of 0.203 and 0.368 nm,
which were assigned to the (111) facet of Ni and the (109) facet Ni Ge O (OH) and Ni Ge O (OH) .
2
flow.
(
3
2
5
4
Fig. 2 (a) UV-vis absorption spectra of Ni/Ni
3
Ge
2
O
5
(OH)
4
and
Ni Ge (OH) . (b) Ni 2p, (c) Ge 3d and (d) O 1s XPS spectra of Ni/
3
2
O
5
4
3
2
5
4
3
2
5
4
of Ni Ge O (OH) , respectively (Fig. S4†). In addition, HR-TEM
3
2
5
4
observation showed that about 10 nm Ni nanoparticles clung
to the Ni Ge (OH) particles (Fig. S4†), corresponding well
with the size of Ni particles from NaBH reduction of Ni(NO )
at 273 K (Fig. S5†).
The scanning electron microscopy (SEM) image showed Fig. 2b shows that the Ni 2p3/2 XPS peaks of Ni/Ni
3
2
O
5
4
The chemical environment of the elements was sub-
sequently analysed by X-ray photoelectron spectroscopy (XPS).
4
3 2
3
Ge
2
O
5
2
(OH)
4
+
that the Ni/Ni Ge O (OH) sample presented an aggregation of at 855.4 and 851.9 eV were respectively assigned to Ni and
3
2
5
4
0
irregular hundred-nanometre spherical particles (Fig. 1b). The Ni . According to the XPS peak area, the percentage compo-
2
+
0
morphology of the Ni
3
Ge
2
O
5
(OH)
4
sample was similar to Ni/ sition of Ni and Ni was respectively calculated to be 84%
Ge (OH) . This suggested that the mole
several-ten nanometer spherical particles (Fig. 1c). For Ni/ ratio of Ni/Ge in Ni/Ni Ge (OH) was about 1.78, which was
Ni Ge (OH) and Ni Ge (OH)
, their low crystallinities in good agreement with the EDS result (Fig. S3†). The Ge
were well confirmed by transmission electron microscopy species was observed at 31.5 eV for Ni/Ni Ge (OH) and at
TEM) observation (Fig. 1d and e). Selected area electron diffr- 32.1 eV for Ni Ge (OH) , respectively (Fig. 2c). About 0.6 eV
action (SAED) patterns revealed an obvious diffraction ring, lower binding energy was probably attributed to the lower crys-
Ni Ge O (OH) and exhibited a denser aggregation of irregular and 16% in Ni/Ni
O
3
2
5
4
3
2
5
4
3
2
O
5
4
4
+
3
2
O
5
4
3
2
O
5
4
O
3
2
5
4
(
3
2
O
5
4
suggesting that Ni/Ni Ge O (OH) and Ni Ge O (OH) were tallinity of Ni/Ni
Ge O (OH) compared to Ni Ge O (OH) , as
2 5 4 3 2 5 4
3
2
5
4
3
2
5
4
3
polycrystalline in nature.
demonstrated in XRD results. Indeed, a higher binding energy
4
+
Fig. 2a shows the UV-vis absorption spectra. As can be seen, of Ge (33.0 eV) was observed due to the high crystallinity of
the Ni Ge O (OH) sample presented a strong UV absorption Ni Ge (OH) (Fig. S7†). In addition, the binding energy of
for a wavelength shorter than 320 nm. Other three absorption reduced Ge in Ni/Ni Ge (OH) was observed at 29.6 eV,
peaks at 350–500, 600–800 and after 900 nm were mainly which was located between 31.5 eV of Ge and 29.0 eV of
O
3
2
5
4
3
2
5
4
δ+
3
2
O
5
4
4
+
2
+ 15
0 19
attributed to the intermediate-band transition of Ni . For Ge . Quantitative analysis revealed that about 9.7% of the
4
+
δ+
3 2 5 4
Ni/Ni Ge O (OH)
, an extraordinary high absorption in the lattice Ge was reduced to Ge during the formation of Ni/
detected wavelength range was observed and overlapped with Ni
the absorption edge of Ni Ge O (OH) . As we know, Ni has a
3
Ge
2 5 4
O (OH) .
During the crystallization of Ni
Ge O (OH) , the Ge–O tetra-
2 5 4
3
2
5
4
3
wide-spectrum range of localized surface plasmonic resonance hedral sheet and the Ni–OH octahedral sheet condensed to
(
LSPR) that enhances the absorption ability and broadens the assemble the tetrahedral–octahedral structural layers, in which
1
6–18
light response range.
Indeed, the as-prepared Ni nano- the hydroxyl in the octahedral sheet was partly substituted for
in NaBH
aqueous the reactive oxygen atom in the tetrahedral sheet accompanied
solution showed a strong light absorption in the wavelength by the formation of Ge–O–Ni chemical bonds (Fig. S8†).
range of 300–1500 nm (Fig. S6a†). And mixing the Ni and Therefore, the O 1s XPS could be deconvoluted into three
Ni Ge (OH)
nanoparticles showed a similar UV-vis absorp- peaks (Fig. 2d). Due to the bigger electronegativity of Ge (2.01)
tion spectrum to Ni/Ni Ge (OH)
(Fig. S6b†), indicating that than Ni (1.91), the peaks at 529.5 and 530.8 eV were attributed
the expanding light absorption mainly originated from the to Ge–O–Ni and Ge–O bonds, respectively. The O 1s XPS
particles by direct reduction of Ni(NO
3
)
2
4
2
0,21
3
2
O
5
4
3
2
O
5
4
2
2
2
3
LSPR effect of Ni nanoparticles. Extra light absorption could peak at 532.0 eV was assigned to the Ni–OH bonds.
enhance the charge carrier production, thus improving the Obviously, compared with Ni Ge (OH) , the mole ratio of
the Ge–O–Ni and Ge–O bonds on Ni/Ni Ge
3
2
O
5
4
CO reduction.
3
2
O
5
(OH)
4
increased,
2
This journal is © The Royal Society of Chemistry 2018
Dalton Trans.

View Article Online
Paper
Dalton Transactions
while the mole ratio of Ni–OH bonds decreased. This evidence located at 1.70 V vs. normal hydrogen electrode (NHE, pH = 7).
indicated that the Ni–OH bonds were slightly unstable and the Therefore, the valence band edge (EVB) of Ni Ge O (OH) was
3
2
5
4
dehydroxylation may occur during the formation of Ni/ 2.30 V vs. NHE, which was powerful enough to oxidize H
Ni Ge (OH) . The decreased amount of Ni–OH bonds would (0.82 V vs. NHE). Combined with the E value (3.88 eV) calcu-
induce the increase in the relative mole ratio of Ge–O–Ni and lated by the UV-vis absorption edge, the conduction band edge
2
O
3
2
O
5
4
g
δ+
Ge–O bonds, although the formation of Ge decreased the (ECB) for Ni
actual mole amount of the coordinated Ge–O bonds. The was able to reduce CO
quantitative FT-IR spectrum analysis (Fig. S9†) revealed that (−0.24 vs. NHE). The EVB and ECB potentials of
(OH) were thermodynamically enough to drive the
reduction to occur.
For Ni/Ni Ge O (OH) , the valence band edge was located at
3
Ge
2
O
5
(OH)
4
remained at about −1.58 V vs. NHE,
2
to form CO (−0.53 V vs. NHE) and CH
4
V
−
1 24
the decreased vibrations of Ge–O (823 cm
)
and Ni–OH Ni
Ge (OH) CO
compared with Ni Ge O (OH) . This fact further suggested
3
Ge
2
O
5
4
bonds (3440 cm−
1 25
)
were both observed for Ni/Ni
O
3
2
5
4
2
3
δ+
2
5
4
3
2
5
4
that the reduced Ge species formed and the partial dehydrox- 0.22 eV below the Fermi level, whereas the valence band
ylation occurred during the formation of Ni/Ni Ge (OH) maximum energy blue-shifted toward the vacuum level at
probably due to the strong reduction ability of NaBH4 approximately −1.65 eV (Fig. S10†). During the formation of
solution. Ni/Ni Ge (OH) , the strong reduction ability of NaBH may
The photocatalytic CO reduction tests were carried out over induce surface disorder and hence produce the mid-gap states.
the as-prepared Ni/Ni Ge O (OH) and Ni Ge O (OH) photo- The disorder-induced mid-gap states could upshift the valence
3
2
O
5
4
,
3
2
O
5
4
4
2
3
2
5
4
3
2
5
4
catalysts using H
typical time–course curves of CO and CH
irradiation by a 300 W Xe lamp for 6 h, the CO and CH yields the flat band potential of the Ni/Ni Ge O (OH) composite was
2
O as a reducing agent. Fig. 3 shows the band maximum energy of Ni
3
Ge
2 5 4
O (OH) , similar to the black
2
6,27
4
2
yields. After full arc TiO .
The Mott–Schottky plot (Fig. S11a†) suggested that
4
3
2
5
4
−
1
(
Fig. 3a and b) were 186.4 and 30.4 μmol gcat.
over Ni/ located at 1.55 V vs. NHE. Therefore, the EVB of Ni/
Ge (OH) was 1.77 vs. NHE. The CB of Ni/
Ni Ge (OH) , respectively, about 3.1 and 6.3 times as much Ni
3
2
O
5
4
3
2
O
5
4
V
E
−
1
as those over Ni Ge O (OH) (59.5 μmol gcat. for CO and Ni Ge O (OH) was expected to downshift due to the for-
3
2
5
4
3
2
5
4
δ+
−
1
4
.8 μmol gcat. for CH
4
). Under a dark reaction or without a mation of Ge , which would be demonstrated in the following
could photocatalytic CO reduction tests under infrared light
only be detected under irradiation, and their yields increased irradiation.
while prolonging the irradiation time, proving that the CO Brunauer–Emmett–Teller (BET) tests showed that the
reduction was a light-driven catalytic reaction. In addition, specific surface area for Ni/Ni Ge (OH) and Ni Ge (OH)
, respectively (Fig. S12a†). The
as a reactant, excluding the effect of possible larger specific surface area resulted from the smaller particle
organic residues on production detection (Table S1†). size of Ni Ge (OH) than Ni/Ni Ge (OH) . The CO
catalyst, no products could be detected. CO and CH
4
2
2
3
2
O
5
4
3
2
O
5
4
2
−1
only O could be detected during vacuum irradiation for 6 h was 16.15 and 33.26 m g
2
cat.
without CO
2
3
2
O
5
4
3
2
O
5
4
2
−
1
The valence band edge of Ni Ge O (OH) was determined adsorbed amount of Ni/Ni Ge O (OH) (3.65 mg gcat. ) was
by valence-band XPS to be 0.60 eV below the Fermi level by the obviously lower than that of Ni
linear extrapolation method (Fig. S10†). The Mott–Schottky (Fig. S12b†). The amount of CO
3
2
5
4
3
2
5
4
−
1
3
2 5 4
Ge O (OH) (8.41 mg gcat. )
adsorbed was normalized by
2
−
2
plot (Fig. S11b†) suggested that Ni Ge O (OH) was a p-type the specific surface area to be 0.25 mg m for Ni Ge O (OH) ,
3
2
5
4
3
2
5
4
conducting semiconductor and the flat band potential was which was much close to 0.23 mg m− for Ni/Ni
This evidence revealed that the CO
the main factor that contributed to the higher CO reduction
2
Ge
O
(OH)
.
3
2
5
4
2
adsorbed capacity was not
2
efficiency over Ni/Ni
To check the effect of light absorption on CO
the CO2 reduction tests were carried out under visible light
λ > 420 nm). As given in Fig. 3c and d, under visible light
irradiation, both Ni/Ni Ge (OH) and Ni Ge (OH)
4
3 2 5 4
Ge O (OH) .
2
reduction,
(
3
2
O
5
4
3
2
O
5
samples showed continuously increased CO and CH4 yields
while increasing the irradiation time. After irradiation for 6 h,
−
1
the CO and CH yields were 55.9 and 20.8 μmol gcat. for Ni/
4
Ni Ge O (OH) , about 17 and 25 times higher than those for
3
2
5
4
−
1
−1
Ni
3
Ge
2
O
5
(OH)
4
(3.1 μmol gcat. CO and 0.8 μmol gcat. CH
yields over Ni Ge (OH) were probably
attributed to the weak visible light absorption via interband
4
).
The low CO and CH
4
3
2
O
5
4
2
+
transition of Ni , as demonstrated by the UV-vis result.
Similarly, an inter-band transition in the defective WO3−x is
2
8
able to drive CO reduction under infrared light irradiation.
2
In addition, steady-state photoluminescence spectra showed
that Ni Ge (OH) exhibited 640 and 770 nm emission peaks
4 3 2 5 4 3 2 5 4
Fig. 3 CO and CH yields for Ni/Ni Ge O (OH) and Ni Ge O (OH)
3
2
O
5
4
under full arc (a and b) and visible light irradiation (c and d) for 6 h.
under the excitation of 470 nm (Fig. S13†), which could be
Dalton Trans.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Dalton Transactions
Paper
ascribed to the recombination from the intermediate band to
2
9
the valence band. For Ni/Ni Ge O (OH) , the broadening
3
2
5
4
emission peak from 620 to 850 nm centered at 750 nm would
correspond to the recombination of abundant defects
3
0
(
Fig. S13†). The decreased PL intensity for Ni/Ni Ge O (OH)
3 2 5 4
3 2 5 4
versus Ni Ge O (OH) could be ascribed to the effective sup-
pression of radiative charge recombination. Additionally, the
average fluorescence lifetimes of Ni/Ni Ge O (OH) and
3
2
5
4
Ni
3
Ge
2
O
5
(OH)
4
were 4.75 ns and 3.40 ns respectively
(Fig. S13†), suggesting the improvement of the efficiency of
charge transport.
In our case, the infrared light response (λ > 800 nm) of
Fig. 4 Schematic energy band diagram for solar CO
into CO and CH
2
overall splitting
Ni
3 2 5 4 2
Ge O (OH) is not able to drive the CO reduction reaction.
−
1
4
.
However, Ni/Ni Ge O (OH) still produced 22.5 μmol g
CO
after infrared irradiation for 6 h
Fig. S14†). This fact indicated that the carbon-based product
3
2
5
4
cat.
and 0.9 μmol gcat.− CH
1
4
(
formation was probably driven by the hot electrons from the
plasma effect of Ni particles. The possible channel of electron
transfer for photocatalysis under infrared irradiation was that
the LSPR excited electrons in plasmonic metal transferred to
the conduction band or mid-gap states of Ni/Ni
3
Ge
2
O
5
(OH)
4
6
2
and subsequently took part in the CO reduction reaction.
2
With the modification of Ni, the resulting hot electrons
increased the charge concentration, thus improving the CO
2
reduction reaction rates. However, no products could be
detected over the physically mixed Ni + Ni Ge O (OH) sample
3
2
5
4
under infrared light irradiation. This evidence meant that Ni/
Ni Ge (OH) was able to drive the half-reaction of water oxi-
dation under infrared light irradiation thermodynamically. As
3
2
O
5
4
a
result, the total CO
2
reduction efficiency over Ni/
Ni
3
Ge (OH) under visible light irradiation was even higher
2
O
5
4
than that over Ni Ge O (OH) under full arc irradiation. The
energy of infrared light (λ > 800 nm) was lower than 1.55 eV,
while the location of the Fermi level of Ni metal was 0.65 V vs.
3
2
5
4
Fig. 5 (a) Ge 3d and (b) O 1s XPS spectra of Ni/Ni
reaction. (c) Ge 3d and (d) O 1s XPS spectra of Ni/Ni
vacuum irradiation for 6 h (denoted as Ni/Ni Ge (OH)
3
Ge
Ge
2
O
5
(OH)
(OH)
4
after
after
3
2
O
5
4
3
2
O
5
4
+ Vac.6 h) and
NHE. Thus, the ECB for Ni/Ni Ge O (OH) was expected to be
3
2
5
4
of Ni/Ni Ge O (OH) + Vac.6 h after CO2 adsorption for 6 h (denoted as
3
2
5
4
downshifted below −0.9 V (NHE), possibly resulting from the Ni/Ni
3
Ge
2
O
5
(OH)
4
2
+ Vac.6 h + CO ads.6 h).
δ+
2+
formation of reduced Ge species. Similar Ge inducing ECB
9
downshift was observed in amorphous α-Zn–Ge–O. Indeed,
the formation of CO and CH
would indicate that the LSPR effect of Ni on Ni/Ni
was energetically enough for driving the CO reduction reac- during photocatalysis was mainly related to the Ge element.
4
under infrared light irradiation
+
31
3
Ge (OH)
2
O
5
4
2
Ni (Fig. S15†). Therefore, the activation of the CO molecule
2
tion to occur, as described in Fig. 4.
However, no obvious change in Ge 3d XPS spectra and Ni 2p
Ge (OH) before and after vacuum
typical Lewis acid, formed on the Ni/ irradiation was observed (Fig. S16 and S17†), further confirm-
. Evidently, after 6 h CO
reduction reaction, ing that the formation of coordinately unsaturated metal sites
the content of Ge further increased and the amount of Ge–O was related to the low crystallinity of Ni/Ni Ge (OH)
bonds decreased (Fig. 5a and b). The increased amount of
Generally, the coordinately unsaturated Ge could serve as
Ge would originate from photogenerated electrons to reduce active sites to interact with O of CO . To check the activation
adsorption under dark conditions was
vacuum irradiation of Ni/Ni Ge O (OH) for 6 h induced about carried out after vacuum irradiation of Ni/Ni Ge (OH)
6.1% increase in the content of Ge species (Fig. 5c and d). Obviously, the quantitative FT-IR analysis indicated that the
XPS results had confirmed that the coordinately unsatu- XPS spectra for Ni
3
2
O
5
4
δ+
rated Ge ,
Ni Ge (OH)
a
3
2
O
5
4
2
δ+
3
2
O
5
4
.
δ+
δ+
2
4
+
9
δ+
Ge , as demonstrated in our previous work. Indeed, the role of Ge , CO
2
O
.
4
3
2
5
4
3
2
5
δ+
1
4
+
−1
The binding energy of Ge decreased from 31.5 to 31.2 eV, intensity of two bands at 1390 and 1560 cm (Fig. 6a), respect-
proving that the coordination environment of surface Ge
changed and the element Ge tended to gain electrons under ing vibrations of OvC–O ,
4
+
ively, corresponding to the symmetric and asymmetric stretch-
−
32,33
both increased after the CO
2
+
irradiation. In addition, no obvious signal of Ni as unsatu- adsorption under dark conditions. Meanwhile, the increased
rated metal sites was detected possibly due to the instability of intensity of the Ge–O band at 823 cm− could be ascribed to
1
This journal is © The Royal Society of Chemistry 2018
Dalton Trans.

View Article Online
Paper
Dalton Transactions
OH species decreased obviously, indicating that Ni–OH could
be consumed by the photogenerated holes. In other words, the
lattice Ni–OH could serve as a proton source. After the H O
2
adsorption (Fig. S19†), the FT-IR peak intensity of Ni–OH
recovered, suggesting that the –OH can be regenerated by H O.
2
The O 1s XPS spectrum (Fig. 5b) and the FT-IR spectrum
(
3 2 5 4
Fig. S20a†) of Ni/Ni Ge O (OH) after reaction showed that
the content of Ni–OH found no obvious drop-off after the
photocatalytic reaction, further confirming the regeneration of
Fig. 6 (a) FT-IR spectra of Ni/Ni
3
Ge
ads.6 h. (b) C 1s XPS spectra of Ni/
after reaction and Ni/Ni Ge (OH) + Vac.6 h + CO
2 5 4
O (OH) + Vac.6 h and Ni/
3 2 5 4
Ni–OH during the reaction. For Ni Ge O (OH) , the Ni–OH
Ni
Ni
3
Ge
Ge
2
O
O
5
(OH)
(OH)
4
+ Vac.6 h + CO
2
species slightly involved with the photocatalytic CO reduction,
3
2
5
4
3
2
O
5
4
2
2
ads.6 h. The inset in (a) showed the enlarged FT-IR spectra for Ni/ as confirmed by the consistent analysis of O 1s XPS spectra
Ni
3
Ge
2
O
5
(OH)
4
.
(Fig. S16b and S16d†) and FT-IR spectra (Fig. S20b†).
Considering that Ni–OH on Ni/Ni Ge O (OH) could be
3
2
5
4
consumed by photogenerated holes, the CO
2
photoreduction
δ+
the interactions of Ge and CO
s XPS spectra of Ni/Ni Ge (OH)
showed that the content of Ge–O at 530.8 eV increased detected under irradiation, confirming that Ni–OH could serve
Fig. 5d). Simultaneously, the Ge 3d XPS spectra demonstrated as the proton source for CO reduction. In the absence of H O,
2
. After CO
2
adsorption, the O tests were performed under full arc irradiation under high-
1
3
2
O
5
4
with vacuum irradiation purity CO2 in the absence of H O. Both CO and CH were
2
4
(
2
2
4
+
that the binding energy of Ge increased from 31.2 to 31.7 eV the CO and CH yields over Ni/Ni Ge O (OH) were 343.2 and
4
3
2
5
4
for Ni/Ni Ge O (OH) with pre-irradiation under vacuum 51.7 μmol gcat.− after irradiation for 6 h (Fig. 7), which were
1
3
2
5
4
before and after CO
2
adsorption, suggesting that the Ge 1.8 times and 1.7 times as much as those over Ni/
tended to lose electrons due to a strong interaction between Ni Ge O (OH) with the addition of H O. The CO reduction
3
2
5
4
2
2
δ+
CO2 and Ge (Fig. 5c). Obviously, electron transfer between efficiency abnormally increased without the addition of H O
2
δ+
unsaturated Ge and CO
The peak at 287.8 eV of the C 1s XPS spectrum for CO
adsorbed Ni/Ni Ge O (OH)
2
occurred.
2 2
as the reactant. Under the coexistence of CO and H O, com-
petitive molecule adsorption would occur during the reduction
with pre-irradiation under reaction of CO by H O. The H O molecule, which is a strongly
vacuum was attributed to the CvO bonds (Fig. 6b), indica- polar molecule, would exhibit a stronger interaction with Ge
2
3
2
5
4
2
2
2
3
4
δ+
1
4,36
tive of the formation of CvO species. During the CO
2
compared to the quadrupole CO molecule.
Therefore,
2
δ+
reduction reaction, the C 1s XPS spectrum could be deconvo- Ge tended to activate H O, resulting in a higher proton
2
luted into three peaks: 287.8 eV for CvO bonds, 286.4 eV for releasing, while in turn the activation of CO was limited, thus
2
3
5
C–O bonds, and 284.6 eV for adventitious standard reference causing low CO and CH4 yields over Ni/Ni Ge O (OH) with
3
2
5
4
carbon (Fig. 6b). The content of CvO bonds at 287.8 eV the addition of H O. Similarly, Ni/Ni Ge O (OH) + Vac.6 h
2
3
2
5
4
decreased and that of C–O bonds at 286.4 eV appeared. This showed lower CO and CH4 yields than those of Ni/
fact meant that CO was activated to the CvO species via a Ni Ge O (OH) (Fig. 7) due to the limitation of proton releas-
strong interaction between O of CO2 and Ge
Ni Ge (OH) . Subsequently, CvO species were converted
2
3
2
5
4
δ+
on Ni/ ing. CO2 reduction without the addition of H O showing
2
3
2
O
5
4
into C–O species via the assistance of photoelectrons, which
may serve as an important intermediate for the formation of
δ+
CO and CH
4
. For Ni
3
Ge
2
O
5
(OH)
4
without Ge , the content of
Ge–O observed from the Ge XPS spectra (Fig. S16†) and the
FT-IR spectra (Fig. S18†) remained unchanged after CO2
adsorption for 6 h, further confirming the activation effect of
δ+
Ge .
The O generated from photocorrosion under vacuum for
2
−
1
1
h was 1472 μmol gcat. for Ni/Ni
gcat. for Ni Ge O (OH) (Table S1†), respectively. This fact
3
Ge
2 5 4
O (OH) and 37 μmol
−
1
3
2
5
4
indicated that the low-crystalline Ni/Ni Ge O (OH) was prone
3
2
5
4
to be photocorroded due to its weak constraint of the crystal
lattice. Under a CO atmosphere, after visible light irradiation
2
−
1
for 1 h the O yield was 897 μmol g
and 14 μmol gcat. for Ni Ge O (OH) , respectively. Obviously,
3 2 5 4
for Ni/Ni Ge O (OH)
3 2 5 4
2
cat.
−
1
the lower O2 generation rate resulted from the interactions
between CO and Lewis acid sites. The quantitative FT-IR ana-
lysis (Fig. 6a) and related O 1s XPS spectra (Fig. 5d) of Ni/
2
Fig. 7 CO (red line) and CH
4
(blue line) yields for Ni/Ni (OH)
3
Ge
2
O
5
4
, Ni/
Ni Ge (OH) + Vac.6 h and Ni/Ni
3
2
O
5
4
3
Ge (OH) without H O as reac-
2
O
5
4
2
Ni Ge O (OH) after vacuum irradiation showed that the Ni– tants under full arc light irradiation for 6 h.
3
2
5
4
Dalton Trans.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Dalton Transactions
Paper
higher CO and CH
4
yields over Ni/Ni
3
GeO
5
(OH)
4
was ascribed releasing from kinetically sluggish water oxidation.
Consequently, the low crystalline Ni/Ni Ge O (OH) exhibits
to the more matching reaction rates.
As a conclusion, proton releasing and CO
equally vital for CO photocatalytic reduction. Ge as a Lewis formance especially compared with Ni
acid activating chemically inert CO2 and Ni as a plasmonic design concept of simultaneously activating CO , promoting
3
2
5
4
2
activation were remarkable visible-light and infrared-light photocatalytic per-
δ+
2
3 2 5 4
Ge O (OH) . The
2
promoter enhancing the light absorption improved CO
reduction performance over Ni/Ni Ge (OH) . Surface Ni–OH can be a new strategy for efficient solar energy utilization. This
as a proton source offered a new path to release protons for work might encourage the study on composite material design
CO reduction. The above experimental results allow us to give to establish the coupling of different modifications of
2
proton releasing and expanding the range of spectral response
3
2
O
5
4
2
deeper insights into the lattice hydroxyl assisted photocatalytic materials.
CO2 reduction and the effect of Ni NPs on photocatalytic
performance.
The possible reaction mechanism for enhanced reduction Conflicts of interest
efficiency of CO over Ni/Ni Ge O (OH) to CO and CH under
irradiation is shown as follows:
2
3
2
5
4
4
There are no conflicts to declare.
þ
ꢀ
Ni=Ni3Ge2O5ðOHÞ þ light ! h þ e
ð1Þ
ð2Þ
4
Acknowledgements
þ
ꢀ
þ
2ꢀ
4
h þ 4Ni ꢀ OH ! 4H þ O2 þ 2Ni ꢀ O
or 4h þ 2H2O ! 4H þ O2
Ge4 þ ð4 ꢀ δÞ eꢀ ! Geδþ
þ
þ
This work was supported primarily by the National Basic
Research Program of China (2013CB632404), the National
Natural Science Foundation of China (51872135, 51572121,
þ
ð3Þ
ð4Þ
21603098, and 21633004), the Natural Science Foundation of
ð4 ꢀ δÞ CO2 þ Geδþ ! ð4 ꢀ δÞ CO2ꢀ þ Ge4þ
Jiangsu Province (BK20151265, BK20151383, and BK20150580),
the program B for outstanding PhD candidate of Nanjing
University (201702B084), the Postdoctoral Science Foundation
of China (2017M611784) and the Fundamental Research
ꢀ
ꢀ
þ
CO2 þ e þ 2H ! CO þ H2O
ð5Þ
ꢀ
ꢀ
þ
or CO2 þ 7e þ 8H ! CH4 þ 2H2O
H2O þ 2Ni–O ꢀ ! 4Ni–OHꢀ
2
2
ð6Þ Funds for the Central Universities (021314380133 and
021314380084).
From the above mechanism, photogenerated electrons to
reduce Ge to form Ge as a Lewis acid and photogenerated
4
+
δ+
holes to oxidize Ni–OH can occur under light irradiation. The
abundant Ge as a Lewis acid is energetically favorable for the
References
δ+
adsorption of CO
another proton source and is regenerated by H O. The CO
2
into CvO species. Ni–OH is used as
1 K. Li, X. An, K. Park, M. Khraisheh and J. Tang, Catal.
Today, 2014, 224, 3–12.
2
2
activation and –OH utilization as a proton source can effec-
tively accelerate this light-driven reaction and achieve excellent
photocatalytic performance. In addition, the LSPR of Ni over
the photocatalyst enables the use of visible light and infrared
light to improve the solar-to-chemical conversion. Therefore,
the design concept for Ni/Ni Ge O (OH) , which is easy to
introduce Ge and possesses adequate –OH and LSPR pro-
perties, provides a novel pathway to enhance photocatalytic
activity.
2 W. Tu, Y. Zhou and Z. Zou, Adv. Mater., 2014, 26, 4607–
4626.
3 S. Wang, S. Yan, X. Ma and J. Gong, Energy Environ. Sci.,
2011, 4, 3805–3819.
4 O. Bikondoa, C. Pang, R. Ithnin, C. Muryn, H. Onishi and
G. Thornton, Nat. Mater., 2006, 5, 189–192.
5 I. Thomann, B. Pinaud, Z. Chen, B. Clemens, T. Jaramillo
and M. Brongersma, Nano Lett., 2011, 11, 3440–3446.
6 T. Woolerton, S. Sheard, E. Reisner, E. Pierce, S. Ragsdale
and F. Armstrong, J. Am. Chem. Soc., 2010, 132, 2132–2133.
3
2
5
4
δ+
7
V. Indrakanti, D. Kubicki and H. Schobert, Energy Environ.
Sci., 2009, 2, 745–758.
Conclusions
8
G. Ramesha, J. Brennecke and P. Kamat, ACS Catal., 2014,
4, 3249–3254.
In summary, the low crystalline hydroxide plasmonic photo-
catalyst with a weak lattice constraint for oxygen atoms is
rationally prepared as a highly efficient catalyst for visible light
9 B. Wang, X. Wang, L. Lu, C. Zhou, Z. Xin, J. Wang, X. Ke,
G. Sheng, S. Yan and Z. Zou, ACS Catal., 2017, 8, 516–525.
or infrared light CO
tremendously expands the range of spectral response. Ge
activates CvO bonds to decrease the energy barriers of CO2
2
reduction. The LSPR of metallic Ni 10 H. Kristoffersen, J. Hansen, U. Martinez, Y. Wei,
δ+
J. Matthiesen, R. Streber, R. Bechstein, E. Lægsgaard,
F. Besenbacher, B. Hammer and S. Wendt, Phys. Rev. Lett.,
2013, 110, 146101.
reduction. The surface lattice hydroxyl is not only joint with
δ+
the formation of Ge but also simultaneously acting as a 11 J. Ding, Z. Dai, F. Qin, H. Zhao, S. Zhao and R. Chen, Appl.
proton source for CO reduction to divert the path of proton
Catal., B, 2017, 205, 281–291.
2
This journal is © The Royal Society of Chemistry 2018
Dalton Trans.

View Article Online
Paper
Dalton Transactions
1
1
1
1
2 H. Li, J. Shang, Z. Ai and L. Zhang, J. Am. Chem. Soc., 2015, 24 S. Rivillon, Y. Chabal, F. Amy and A. Kahn, Appl. Phys. Lett.,
37, 6393–6399. 2005, 87, 253101.
3 H. Li, J. Shang, H. Zhu, Z. Yang, Z. Ai and L. Zhang, ACS 25 S. Yang, X. Wu, C. Chen, H. Dong, W. Hu and X. Wang,
Catal., 2016, 6, 8276–8285. Chem. Commun., 2012, 48, 2773.
4 L. Lu, B. Wang, S. Wang, Z. Shi, S. Yan and Z. Zou, Adv. 26 X. Chen, L. Liu, Y. Peter and S. Mao, Science, 2011, 331,
Funct. Mater., 2017, 27, 1702447. 746–750.
5 N. Kuleshov, V. Shcherbitsky, V. Mikhailov, S. Kück, 27 L. Shi, Z. Li, T. Dao, T. Nagao and Y. Yang, J. Mater. Chem.
1
J. Koetke, K. Petermann and G. Huber, J. Lumin., 1997, 71,
65.
6 X. Meng, T. Wang, L. Liu, S. Ouyang, P. Li, H. Hu, T. Kako,
A, 2018, 6, 12978–12984.
28 L. Liang, X. Li, Y. Sun, Y. Tan, X. Jiao, H. Ju, Z. Qi, J. Zhu
and Y. Xie, Joule, 2018, 2, 1004.
2
1
1
H. Iwai, A. Tanaka and J. Ye, Angew. Chem., Int. Ed., 2014, 29 C. Yang, M. Qin, Y. Wang, D. Wan, F. Huang and J. Lin, Sci.
26, 11662. Rep., 2013, 3, 1286.
7 Z. Xiong, X. Chen, X. Wang, L. Peng, D. Yan, H. Lei, Y. Fu, 30 L. Ye, K. Deng, F. Xu, L. Tian, T. Peng and L. Zan, Phys.
1
J. Wu, Z. Li, X. An and W. Wu, Appl. Surf. Sci., 2013, 268,
24.
8 H. Liu, T. Dao, L. Liu, X. Meng, T. Nagao and J. Ye, Appl.
Chem. Chem. Phys., 2012, 14, 82–85.
31 L. Van Robbroeck, E. Goovaerts and D. Schoemaker, Phys.
Status Solidi B, 1985, 132, 179–187.
5
1
1
2
2
2
2
Catal., B, 2017, 209, 183.
9 R. Shalvoy, G. Fisher and P. Stiles, Phys. Rev. B: Solid State,
32 V. Nadtochenko, A. Rincon, S. Stanca and J. Kiwi,
J. Photochem. Photobiol., A, 2005, 169, 131.
33 K. Ogura, M. Kobayashi, M. Nakayama and Y. Miho,
J. Electroanal. Chem., 1998, 449, 101.
34 Y. Wu, P. Wang, X. Zhu, Q. Zhang, Z. Wang, Y. Liu, G. Zou,
Y. Dai, M. Whangbo and B. Huang, Adv. Mater., 2018, 30,
1704342.
1
977, 15, 1680.
0 A. Krasilin and E. Khrapova, Russ. J. Appl. Chem., 2017, 90,
2.
2
1 Y. Yang, Q. Liang, J. Li, Y. Zhuang, Y. He, B. Bai and
X. Wang, Nano Res., 2011, 4, 882.
2 K. Li and D. Xue, J. Phys. Chem. A, 2006, 110, 35 J. Hou, H. Cheng, O. Takeda and H. Zhu, Angew. Chem., Int.
1332. Ed., 2015, 127, 8600.
3 I. Casella, M. Guascito and M. Sannazzaro, J. Electroanal. 36 F. Brandani and D. Ruthven, Ind. Eng. Chem. Res., 2004, 43,
Chem., 1999, 462, 202. 8339.
1
Dalton Trans.
This journal is © The Royal Society of Chemistry 2018