CdS nanorods anchored with CoS2 nanoparticles for enhanced photocatalytic hydrogen production

Article abstract of DOI:10.1016/j.apcata.2019.117281

Herein, we report the use of cobalt sulfide (CoS₂) as efficient and inexpensive co-catalyst of CdS nanorods for photocatalytic water splitting. The aim is to explore the use of earth-abundant cobalt species to replace precious metals for the photocatalytic reactions. The results of first-principles DFT simulation and planar-averaged differential charge density calculation reveal that at the CoS₂/CdS interface, CoS₂ has zero band gap which is a class nature of precious metals, and functions as electron trap to enhance the transfer of hot electrons from CdS to CoS₂. Owing to the merits of efficient charge separation, high exposure of active sites as well as large surface area, the CoS₂/CdS composites exhibit outstanding photocatalytic activity in H₂ production under visible light (?58 mmol·g^?1 h^?1), which is about 19 times that of CdS nanorods alone and 3 times that of 1 wt%Pt/CdS under the same conditions.

Full text of DOI:10.1016/j.apcata.2019.117281

Applied Catalysis A, General 588 (2019) 117281
Contents lists available at ScienceDirect
Applied Catalysis A, General
journal homepage: www.elsevier.com/locate/apcata
CdS nanorods anchored with CoS nanoparticles for enhanced
2
photocatalytic hydrogen production
Jie Tanga,1, Bin Gaoa,1, Jinbo Pan , Lang Chen ⁎, Zihao Zhao , Sheng Shen , Jun-Kang Guo ⁎,
a,1
a,
a
a
a,
b
a,
Chak-Tong Au , Shuang-Feng Yin ⁎
a
State Key Laboratory of Chemo/Biosensing and Chemometrics, College of Chemistry and Chemical Engineering, Provincial Hunan Key Laboratory for Cost-effective
Utilization of Fossil Fuel Aimed at Reducing Carbon-dioxide Emissions, Hunan University, Changsha, 410082, Hunan, People’s Republic of China
b
College of Chemistry and Chemical Engineering, Hunan Institute of Engineering, Xiangtan, 411104, Hunan, People’s Republic of China
A R T I C L E I N F O
A B S T R A C T
Keywords:
2
Herein, we report the use of cobalt sulfide (CoS ) as efficient and inexpensive co-catalyst of CdS nanorods for
CoS
Photocatalysis
DFT calculation
2
/CdS composites
photocatalytic water splitting. The aim is to explore the use of earth-abundant cobalt species to replace precious
metals for the photocatalytic reactions. The results of first-principles DFT simulation and planar-averaged dif-
ferential charge density calculation reveal that at the CoS
nature of precious metals, and functions as electron trap to enhance the transfer of hot electrons from CdS to
CoS . Owing to the merits of efficient charge separation, high exposure of active sites as well as large surface
area, the CoS /CdS composites exhibit outstanding photocatalytic activity in H production under visible light
), which is about 19 times that of CdS nanorods alone and 3 times that of 1 wt%Pt/CdS under
the same conditions.
2 2
/CdS interface, CoS has zero band gap which is a class
H
2
production
2
2
2
−
1
−1
(
˜58 mmol·g
h
1
. Introduction
cheaper [4], such as transition-metal sulfides (e.g., WS
23–25], NiS [26,27], Co [28,29], MnS [30]), and many have
showed excellent performance in a majority of photocatalytic systems.
In our previous works, WS and MoS nanoplates were anchored on CdS
nanorods, and there was great improvement in hydrogen production
under visible light irradiation [31–33]. Recently, there were reports on
2 2
[22], MoS
[
x y
S
With the ever increasing demand of energy, the burning of fossil
fuels is inevitable and environmental pollution is a global concern.
Extensive studies have been conducted in search of energy resources
that are environment-friendly and renewable [1–4]. In this aspect,
hydrogen is the most commendable. The photocatalytic generation of
hydrogen from water splitting using inexhaustible solar energy is a
state-of-the-art strategy, and the use of an efficient photocatalyst is
critical [5–7]. Many attempts have been made to develop high-effi-
ciency photocatalysts, such as phase and morphology adjustment [8],
doping with metal and non-metal elements [9,10], fabrication of het-
erojunctions [11,12] and the use of co-catalysts [13–16]. Among them,
the introduction of a co-catalyst can not only suppress the recombina-
tion of charge carriers through electron trapping but also provide active
sites for proton reduction. Noble metals such as Pt, Pd, Au and Ag have
been widely used as efficient co-catalysts for hydrogen production due
to their low Fermi levels [17–21]. However, the high price and rela-
tively rare resource of precious metals hampers their usage. Therefore,
it is extremely urgent to explore alternatives that are economically
commendable.
2
2
the coupling of cobalt sulfide (e.g., CoS [34], CoS
2 3 4
[35], Co S [36],
Co [28,29]) with semiconductors to enhance the separation of
9 8
S
charge carriers for the promotion of photocatalytic activity. For ex-
ample, through rational design of Z-scheme system, Tan et al. devel-
oped hierarchical CdS-supported Co S nanoparticles that displayed
9 8
high activity and stability in photocatalytic hydrogrn evolution (PHE).
Furthermore, over CdS nanorods modified by cobalt-based co-catalysts
2 3 4
(Co(OH) , Co O , CoS, CoO), Lang et al. observed high photocatalytic
activity and reported that the cobalt-containing species play a sig-
nificant role in the trapping and transport of hot electrons. According to
these photocatalytic systems, cobalt sulfide as a co-catalyst can
strengthen the absorption of solar light, accelerate the separation of
photogenerated charge carriers, and offer ample active sites for reac-
tions [37]. Hou and coworkers reported a hydrogen production effi-
−
1 −1
ciency of 1232 μmol·h
g
2 3 4
over CoS /g-C N hybrid composites [35].
Co-catalysts that are based on earth abundant materials are much
Fu et al. improved the light harvesting ability of carbon quantum dots
⁎
Corresponding authors.
E-mail addresses: huagong042cl@163.com (L. Chen), jkguozd@126.com (J.-K. Guo), sf_yin@hnu.edu.cn (S.-F. Yin).
The authors contributed equally to this work.
1
https://doi.org/10.1016/j.apcata.2019.117281
Received 21 August 2019; Received in revised form 29 September 2019; Accepted 3 October 2019
Available online 08 October 2019
0926-860X/ © 2019 Elsevier B.V. All rights reserved.

J. Tang, et al.
Applied Catalysis A, General 588 (2019) 117281
(
CDs) by adding CoMo-S in the photocatalytic system [38]. Reddy and
and ethanol for three times, and dried at 80 °C overnight. Altogether
coworkers also used C@CoS as co-catalyst to enhance the PHE activity
2
seven CoS
1:3, 1:4, 1:5, and 1:6 were synthesized, and are herein denoted as
CC0.5, CC1, CC2, CC3, CC4, CC5, and CC6, respectively. The pure CoS
catalyst was synthesized following the above steps but without the
addition of CdS.
2
/CdS composites with Co:Cd molar ratio of 1:0.5, 1:1, 1:2,
of CdS nanorods [37]. It is hence envisaged that similar to noble metals,
cobalt sulfides may facilitate the transfer of photogenerated charge
carriers at the heterojunctions.
2
It was reported that cobalt disulfide (CoS
conductivity (6.7 × 10 S cm
2
) with intrinsic metallic
at 300 K) displayed remarkable elec-
trocatalytic and photoelectrocatalytic activities [39]. Using metallic
cobalt pyrite CoS (micro- and nanostructure) as catalysts, Matthew
3
−1
2
.1.3. Synthesis of physically mixed CoS
For comparison purpose, a sample of CoS
prepared by physical mixing (Co:Cd = 1:4, denoted as CC4(P)). Briefly,
mmol of as-prepared CoS powder and 4 mmol of CdS nanorods were
2
-CdS composite
2
2
-CdS composite was
et al. observed good electrocatalytic performance for hydrogen evolu-
tion reaction (HER) [40]. Using Si microwires surrounded by cobalt
disulfide as photocathode for solar hydrogen evolution, Chen and
1
2
thoroughly ground in an agate mortar for 30 min, and the collected
powder was dried at 80 °C overnight.
coworkers found that CoS
as a co-catalyst to reduce the recombination of charge carriers [41]. In
other words, CoS can promote the migration and separation efficiency
2
not only served as protective layer, but also
2
2.1.4. Synthesis of Pt/CdS composite
of photoelectrons. Despite the presence of cobalt sulfide is known to be
beneficial for electron transfer from semiconductors, the exact nature of
charge separation at the composite interfaces is always neglected.
The Pt/CdS composite was prepared by loading 1 wt% of Pt on CdS
according to the photoreduction method reported by Zhang et al. [43].
First, 50 mL of H PtCl ·6H O solution containing 2 mg of Pt was pre-
2
6
2
In this work, we facilely anchored CoS
norods by an in-situ solvothermal method, and analyzed the band gap
as well as the shift direction of interfacial electrons at the CoS /CdS
2
nanoparticles on CdS na-
pared. Then 200 mg of CdS nanorods was added to the solution, and the
mixture was allowed to stand at 25 °C for 2 h. The solution was then
illuminated using a 300 W Xe lamp (with a cut-off filter to remove light
with λ < 400 nm) for 2 h with stirring. Finally the yellow powder was
collected, washed with deionized water and absolute ethyl alcohol for
three times, and dried at 80 °C overnight.
2
heterojunction by DFT calculations. Owing to large surface area, high
exposure of active sites, and effective transfer of photogenerated elec-
trons from CdS to CoS
2
, the CoS
2
/CdS composite gave a H
2
production
−
1
−1
rate of ˜58 mmol·g
h
(λ ≥ 400 nm), much higher than that of CdS
or Pt/CdS.
2.2. Characterization
2
. Experimental section
Powder X-ray diffraction (XRD, D/MAX-2000/PC, λ = 0.15406 nm,
Rigaku Corporation) was employed for the characterization of crystal
structure. The morphology and elemental mapping analyses were
conducted over a field emission scanning electron microscope (FE-SEM)
2
2
.1. Preparation
.1.1. Synthesis of CdS nanorods
(
Hitachi S-4800) and a transmission electron microscope (TEM, accel-
erating voltage: 100 kV, Tecnai G2 F20). The element contents of
samples were calculated according to the results of ICP-OES analysis
Varian 720). The surface composition and chemical states of elements
was analyzed by means of X-ray photoelectron spectroscopy (XPS,
CdS nanorods were prepared according to the solvothermal method
described by Jiang et al. [42]. Typically, 20.23 mmol of CdCl
and 60.69 mmol of CH S were dissolved in ethanediamine (60 mL).
Then the solution was transferred to a polyphenylene-lined autoclave of
00 mL and heated at 160 °C for 50 h. The yellow product was collected
2 2
·2.5H O
4 2
N
(
1
Thermo SCIENTIFIC K-Alpha). The specific surface area (BET) of sam-
by filtration, washed with de-ionized water and ethanol for three times,
and dried at 80 °C for 12 h.
ples was obtained by using
a Gemini VII 2390 instrument
(
Micromeritics Instrument Corp.). UV–vis diffuse reflectance spectra of
samples were recorded using a spectrophotometer (Cary 100, Agilent).
Photoluminescence (PL) measurement was performed on a fluorescence
spectrophotometer (excitation light: 350 nm). An electrochemical ana-
lyzer (CHI660D, Shanghai Chenhua) was employed to record photo-
electrochemical activity, using Ag/AgCl (3 M KCl) as reference elec-
trode, Pt wire as counter electrode, prepared sample (20 mg powder,
0.5 mL Nafion, 1.5 mL alcohol, and dried at 80 °C for 2 h) as working
2
.1.2. Synthesis of CoS
The synthetic process of CoS
Scheme 1. Briefly, 1 mmol of Co(NO
2
/CdS composites
/CdS composites is illustrated in
·6H O and 2 mmol of thiourea
2
3
)
2
2
were dissolved in 80 mL of ethylene glycol, and the mixture was stirred
for 10 min to form a pink solution. Then a certain amount of the as-
prepared CdS nanorods was added with stirring in a period of 10 min.
The mixture was then subject to ultrasonic treatment with stirring for
2 4
electrode, and sodium sulfate (Na SO , 0.2 M) as electrolyte. The data
6
0 min. The as-generated yellow solution was transferred into a poly-
for the plotting of M-S curves were measured using sodium sulfate
electrolyte solution (0.2 M, pH = 7), and the amplitude voltage was
5 mV at a selected frequency. The final potentials vs. Ag/AgCl were
phenylene-lined autoclave of 100 mL and heated at 180 °C for 24 h. The
final product was collected by filtration, washed with deionized water
2
Scheme 1. Synthetic procedure of CoS /CdS composites.
2

J. Tang, et al.
Applied Catalysis A, General 588 (2019) 117281
translated into the standard hydrogen electrode using the below
equation [44]:
CoS
2
/CdS samples are displayed in Fig. 1b, which indicates the pre-
sence of Cd and S on CdS, Co and S on CoS
CdS composites. The Co 2p spectrum of CoS
at 797.6, 793.9, 781.5, and 778.7 eV as well as two shake-up satellite
peaks (Fig. 1c). The peaks at 797.6 eV and 781.5 eV are Co 2p1/2 and
2
, and Co, Cd, S on the CoS
2
/
2
shows deconvoluted peaks
E
NHE = EAg/AgCl + 0.197
We carried out the density functional theory (DFT) calculations with
the Perdew-Burke-Ernzerhof (PBE) exchange-correlation fuctional in
the Vienna ab initio simulation package (VASP) and employed the
empirical correction method proposed by Grimme (DFT-D2) to describe
the van deer Waals (vdW) interaction. The Heyd-Scuseria-Ernzerhof
2
2p3/2 signals of CoS , while those at 793.9 eV and 778.7 eV are Co 2p1/2
and 2p3/2 signals of cobalt oxide [46,47]. As for CdS, it shows Cd 3d
peaks at 410.8 eV and 404.1 eV (Fig. 1d), and S 2p peaks at 161.5 eV
and 160.4 eV (Fig. 1e), in agreement with literature data [31,42].
Compared with the S 2p spectrum of CdS, that of CoS
plicated (Fig. 1e). The CoS peaks at 163.2 eV and 161.9 eV are as-
cribable to the S 2p1/2 and 2p3/2 signals of CoS , which overlap strongly
2
is more com-
(
HSE06) hybrid functional was adopted to investigate the band gap due
to the underestimation of PBE. The energy cutoff for plane-wave basis
set and the Monkhorst-Pack k-point grid were set as 350 eV and 3 × 3
2
2
×
1, respectively. A vacuum space of over 15 Å along the c axis was
with those of SeC species. As for the peak at 168.9 eV, it is attributable
to the S 2p1/2 and 2p3/2 signals of SeO entities [38,47,48]. The XPS
spectra of the hybrid CoS /CdS samples (i.e. CC2, CC4 and CC6) are
2
rather similar (Fig. 1f,g,h). The Cd 3d3/2 and Cd 3d5/2 peaks are at
411.6 eV and 404.8 eV, respectively (Fig. 1d), which are consistent with
the Cd2+ signals of CdS in literature [13,31]. The Co 2p signals at
constructed to minimize the interaction between adjacent layers. The
geometric structure was optimized until the convergence standard of
total energy was 1.0 × 10 eV and the force on per atom was smaller
–
5
–
1
than 0.05 eV∙Å
.
2.3. Photocatalytic hydrogen generation
796.6 eV and 780.6 eV are assignable to Co-S of CoS
2
, while those at
794.0 eV and 778.9 eV) to Co-O of cobalt oxide, and the two peaks in
Photocatalytic hydrogen production experiments were carried out
purple color are attributed to shake-up satellites [45,47]. In Fig. 1h, the
peaks at 161.1 eV and 163.0 eV are ascribed to the S 2p3/2 and 2p1/2 of
Co-S bond, and the signals at 168.8 eV was caused by SeO bond, which
is in agreement with previous reports [29,43]. Compared to CdS, CC4
exhibits positive shift (˜0.8 eV) in Cd 3d and S 2p binding energy, while
in a sealed quartz reactor (250 mL) equipped with pumping facility. The
reactor top was illuminated with a 300 W Xe lamp (HSX-F300,
Perfectlight, Beijing) with a cut-off filter (λ ≥ 400 nm). First, 20 mg of
photocatalyst was put in the reactor lumen together with an aqueous
solution containing lactic acid or Na
2
S/Na
2
SO
3
(50 mL, 20 vol%). Then
compared to CoS
binding energy, which indicate electron transfer from CdS to CoS
[28,49,50]. Based on the XPS results, it is deduced that there are close
interaction between CdS and CoS , forming interfaces in CC2, CC4 and
2
, there is negative shift (˜0.8 eV) of Co 2p and S 2p
the gases were removed using a vacuum pump, and the reaction system
was kept in the dark with constant stirring for 30 min to establish ad-
sorption-and-desorption equilibrium. The production of hydrogen was
hourly monitored using a gas chromatograph (Agilent, 7820A, 5A
molecular sieve column) equipped with a thermal conductivity de-
tector. The apparent quantum yield (A.Q.Y.) was measured upon
monochromatic light irradiation according to the method reported in
our previous work [31] using the following equations:
2
2
CC6 that act as ‘electron bridges’ to accelerate charge transfer.
According to the XRD results, the most exposed surfaces of the
CoS
(Fig. 1a). We hence conducted the first-principles DFT simulation
(Fig. 2a and b) based on the structures of CoS (200) and CdS (001)
surfaces: cubic phase with Pa-3 space group for CoS and hexagonal
phase with P63mc space group for CdS. The geometric structure of
CoS /CdS is illustrated in Fig. 2c, and the lattice mismatch is no more
than 2.53%. The calculated binding energy of CoS /CdS is less than
zero, which indicates that CoS and CdS can form a heterostructure that
2 2
/CdS composites are those of the CoS (200) and CdS (001) planes
2
2
number of reacted electrons
number of incident photons
A.Q. Y. =
× 100%
2
2
2
× (number of evolved hydrogen molecules)
=
× 100%
2
number of incident photons
is thermodynamically stable. We performed first-principle calculation
to explore the innate characteristics of charge transfer and separation.
3. Results and discussion
The Fermi levels (E
F 2
) of the exposed CdS (001) and CoS (200) surfaces
are estimated as follows:
The XRD patterns of CdS, CoS
2
and CoS
2
/CdS samples are compiled
Φ = Evac − E
F
in Fig. 1a. The peaks of CdS nanorods match well with those of hex-
agonal phase CdS (JCPDS 41-1049) [42], while the main peaks of CoS
nanoparticles are assignable to those of cubic CoS (JCPDS 41-1471)
45], suggesting that pure CdS and CoS were successfully prepared. As
for the CoS /CdS composites, the characteristic peaks of CdS can be
detected but those of CoS cannot. The absence of CoS signal could be
2
Where Φ is the work function, which is calculated to be 5.17 eV and
5.23 eV, respectively, for CdS (001) and CoS
2
2
(200) surfaces
[
2
(Fig. 2d − e), and Evac is the energy of a stationary electron at the va-
cuum level (assumed to be 0 eV). Accordingly, the E of CdS (001) and
CoS (200) surfaces can be calculated to be -5.17 and -5.23 eV, re-
spectively. The results are consistent with the work of Xu et al., which
revealed that the work function (vs. vacuum level) of CdS and CoS was
5.18 eV and 5.49 eV, respectively [51]. It is obvious that the E of CoS
(200) surface is more negative than that of CdS (001) surface. Thus,
when the CdS (001) and CoS (200) surfaces are in contact, electrons
tend to migrate from CdS to CoS for potential equilibrium, restraining
the recombination of charge carriers as a result. To further demonstrate
the charge transfer at the CoS /CdS interface, the planar-averaged
2
F
2
2
2
due to its high dispersion, low content, and/or poor crystallinity. Si-
milar phenomena were reported in the works of Chen et al. [13] and
Jiang et al. [42]. It is worth pointing out that the relative intensity of
the (100) peak (2θ = 24.807°) of CdS as well as that of the composites
are much higher than that of standard CdS (in the PDF card), indicating
the high exposure of the CdS (100) plane. Similarly, the XRD pattern of
2
F
2
2
2
CoS
2
indicates the high exposure of the (200) plane. Furthermore, the
/CdS composites was verified by ICP-
actual composition of the CoS
2
2
OES analysis (Table S1, supporting information, SI†), and the genuine
Co:Cd molar ratios are quite in agreement with the nominal ones.
Overall, the results of XRD and ICP analyses evidence the successful
differential charge density was calculated and the result is shown in Fig.
S1 (SI†). The positive value represents charge accumulation, while the
negative value indicates charge depletion. It is apparent that at the
fabrication of the CoS
no detection of impurity phases, it is deduced that the coupling of CoS
2
/CdS composites. Furthermore, because there is
CoS
the CoS
2
/CdS interface, electrons primarily transfer from the CdS layer to
layer.
2
2
with CdS has little effect on the crystallinity and phase structure of CdS.
We performed XPS analysis to investigate the chemical states and
2
composition of surface elements. The survey spectra of CdS, CoS and
As shown in Fig. 3a and b, the SEM images of pure CdS clearly show
nanorods with an average length of 800 nm. The surface of the na-
norods is smooth, showing no impurity or pores [44]. On the other
3

J. Tang, et al.
Applied Catalysis A, General 588 (2019) 117281
Fig. 1. (a) XRD patterns, (b) XPS survey spectra, and high-resolution XPS spectra of samples: (c) (f) Cd 3d, (d) (g) S 2p, (e) (h) Co 2p.
hands, pure CoS
2
exists as microparticles approximately 1.5 μm in size
2
0.24 nm are attributed to the (210) plane of CoS and (102) plane of
(
Fig. 3c and d), which are assemblies of nanoparticles. The morphology
CdS, respectively. The results are in line with those of previous reports
[13,53,54]. The elemental mapping (Fig. 4e − h) and EDX analysis
results (Fig. S3, SI†) of CC4 further disclose homogeneous distribution
of CoS
the CoS
2
/CdS heterostructure is displayed in Fig. 3e and f. It is clear that
nanoparticles (about 15 nm in size) are evenly distributed on
2
the CdS nanorods with no sight of aggregation, and hence could not be
detected in XRD analysis. Furthermore, the BET surface area of CdS and
CC4 is 32.7 m g and 34.7 m g , respectively (Fig. S2, SI†). Despite
the difference in specific surface area is not apparent, there are much
of Cd, S and Co elements which confirms the high dispersion of CoS
2
nanoparticles on CdS.
The absorption coefficient of the composites is determined ac-
cording to the following equation:
2
−1
2 −1
2
more active sites on the latter due to the presence of CoS .
1
2
2
The TEM and HRTEM images of pristine CdS nanorods and CC4 are
displayed in Fig. 4. In consistent with the SEM results, the TEM image of
Fig. 4a clearly shows that the surface of CdS nanorods is smooth. The
HRTEM image (Fig. 4b) reveals crystal lattice fringes of 0.36 nm which
is assignable to the (100) lattice plane of CdS [52]. The images of CC4
in Figs. 4c and d exhibit the presence of particles approximately 15 nm
in size on the CdS nanorod, and their presence does not alter the
structure of the CdS nanorod. The lattice spacings of 0.25 nm and
α (ω) = ( 2 )ω[ ε1(ω) + ε2 (ω) − ε1(ω)]2
Where
1 2
ε (ω) and ε (ω) are the real and imaginary parts of dielectric
function, respectively, and ω is defined as the function of optical fre-
quency. As shown in Fig. 5a, CdS has poor optical absorption and needs
improvement in the 1.6 − 3.1 eV region for visible light absorption,
while the absorption coefficient of CoS
Upon coupling of CdS and CoS , there is significant enhancement of
light absorption. In order to disclose the energy band structure of CdS
2
is higher than that of CdS.
2
4

J. Tang, et al.
Applied Catalysis A, General 588 (2019) 117281
2 2 2
Fig. 2. Side and top views of (a) CoS , (b) CdS and (c) CoS /CdS; calculated electrostatic potentials along z direction of (d) CdS and (e) CoS .
and CoS
DFT). The electronic band structure of CoS
Fig. 5b and c. It is clear that the band gap of CoS
consistent with the result of Chen et al. [55]. Meanwhile, the band gap
of CdS is 1.52 eV, in agreement with the calculation result of Garg et al.
that planar CdS is a semiconductor with a band gap of 1.60 eV [56]. The
agreement indicates that our calculation adopting the PBE approach is
2
, we carried out calculation using density functional theory
and CdS are shown in
is 0 eV, which is in
Mott–Schottky curves were plotted to qualitatively analyze the p–n
junctions in the absence of illumination as illustrated in Fig. S4 (SI†).
(
2
−2
−2
2
The C /F -E plot of CdS displays a positive slope, which is char-
acteristic of n-type semiconductor. By obtaining the X intercept by
means of extrapolation of Mott-Schottky plots (in supporting informa-
tion), the flat-band potential (Vfb) vs. NHE of CdS at pH = 7.0 is
roughly -0.719 V. Using the band gap of CdS (2.38 eV) estimated from
the UV–vis DRS graph, the valence band of CdS is found to be +1.661 V
appropriate. The outcomes (Fig. 5) predict that CoS
harvesting ability and good photoelectron conductivity (Eg = 0 eV at
the CoS /CdS interface) at the CoS /CdS heterojunction. It is hence
theoretically confirmed that the metallic property of CoS is crucial in
2
has good light-
vs. NHE at pH = 7.0. With built-in electric field at the CoS
erojunction, the photoelectrons quickly transfer from CdS to CoS
leading to enhanced H production under visible light irradiation.
The photoluminescence (PL) spectra of CoS /CdS composites are
2
/CdS het-
2
2
2
,
2
2
the transfer and separation of photoelectrons for photocatalytic evolu-
tion of hydrogen in water splitting.
The UV–vis diffuse reflectance spectra of the as-prepared samples
are depicted in Fig. 6a. Pure CdS nanorods show obvious absorption in
2
shown in Fig. 6c. That of CdS shows an apparent peak at about 537 nm
due to the direct recombination of photogenerated charge carriers. A
lower PL signal means higher separation efficiency of photo-excited
the region of 400–530 nm as previously reported [15,42], and CoS
exhibits absorption that covers almost the entire visible light region
45]. As for the composites, they show light absorption pattern similar
2
charge carriers. Apparently, the loading of CoS
CdS nanorods results in lowering of PL intensity. Among the CoS
composites, CC4 shows the lowest PL intensity and hence is the most
2
nanoparticles on the
/CdS
2
[
to that of CdS but with enhanced absorption in the 530–700 nm region,
efficient in preventing the recombination of charge carriers. To further
in agreement with the DFT calculation results (Fig. 5a). The phenom-
2
confirm the mobility of photogenerated charge carriers in the CoS /CdS
enon could be ascribed to the strong absorption capacity of CoS
visible light region and the interfacial interaction between CdS and
CoS . In addition, the band gaps of the CoS /CdS samples can be ob-
tained using the following equation [15]:
2
in
samples, transient photocurrent was recorded and analyzed using
electrodes coated with the samples under visible light irradiation
(Fig. 6d). It is clear that photocurrent could only be obtained upon light
irradiation. Specifically, pure CdS nanorods is the lowest in photo-
current intensity while CC4 the highest, giving an intensity order of
CdS < CC2 < CC6 < CC4. Again, the results demonstrate that the
2
2
αhv = A(hv-E
g
)n/2
2
loading of CoS nanoparticles on CdS nanorods can efficiently enhance
the transfer and separation of photoexcited electrons and holes.
Photocatalytic hydrogen evolution under visible light illumination
where α, h, v, A and E
g
stand for the absorption coefficient, Planck’s
constant, light frequency, proportionality constant, band gap, respec-
tively; and n is dependent on the type of semiconductor. According to
the Kubelka–Munk plots in Fig. 6b, the band gap values of CdS and CC4
are 2.38 eV and 2.37 eV, respectively, which are larger than the theo-
retical value (1.52 eV). It is plausibly because the former are experi-
mental values based on bulk samples while the latter is calculated value
was used to assess the photocatalytic performance of the CoS
composites. As illustrated in Fig. 7a, there is no detection of H
CoS
the activity of CdS is limited (3.1 mmol·g
3 h). Gratifyingly, all the CoS /CdS composites show activity higher
2
/CdS
2
over
2
nanoparticles owing to its class nature of precious metals, while
−1
−1
h
after an irradiation of
based on a small model of layered CdS at the CoS
2
/CdS interface. The
2
5

J. Tang, et al.
Applied Catalysis A, General 588 (2019) 117281
2
Fig. 3. SEM images of (a) (b) CdS nanorods, (c) (d) CoS and (e) (f) CC4.
Fig. 4. TEM images of (a) (b) CdS nanorods, (c) (d) CC4 and (e)–(h) the relevant EDX mappings of Cd, S, Co.
6

J. Tang, et al.
Applied Catalysis A, General 588 (2019) 117281
Fig. 5. (a) The optical absorption spectra of
CoS
2
,
CdS and CoS
2
/CdS; electronic band
structures of (b) CoS
2
and (c) CdS calculated by
PBE functional. The Fermi level is placed at zero
and marked by horizontal green dotted lines.
The vertical red dash lines G, M and K represent
the high-symmetry k points: (0, 0, 0), (0, 1/2, 0)
and (-1/3, 2/3, 0), respectively. (For inter-
pretation of the references to colour in this
figure legend, the reader is referred to the web
version of this article).
than that of CdS (Fig. 7b). Among the prepared catalysts, CC4 is the
there is apparent decrease in H
2
production across the four cycles.
S/0.25 M Na SO ) is used,
−
1
−1
most efficient, giving a H
2
production rate of 58.1 mmol·g
h
,
Nonetheless, when Na S/Na SO (0.35 M Na
2
2
3
2
2
3
which is ca. 19 times that of CdS nanorods, four times that of CC4(P)
there is no apparent loss, confirming the high stability of CC4 in alka-
line solution. We employed the XRD, TEM and XPS techniques to
−
1
−1
prepared by physical mixing (14.7 mmol·g
h
), and three times that
−1
−1
of 1 wt%Pt/CdS (19.8 mmol·g
h
). The results reveal that CoS can
2
analyze CC4 before and after reaction using either lactic acid or Na
2
S/
replace noble metals as an effective co-catalyst for photocatalytic gen-
Na SO as sacrificial agent. The XRD patterns suggest no change of
2
3
eration of hydrogen in water splitting. Among the CdS-based photo-
catalysts using Co-containing compounds as co-catalysts for H pro-
2
duction under visible light irradiation, CC4 is moderate in activity.
However, among the reported CdS photocatalysts modified by cobalt
sulfides, CC4 is the highest in activity as well as apparent quantum yield
phase structures after reaction (Fig. S6, SI†). There is no change in the
Cd 3d spectra neither after reaction (Fig S7, SI†). Hence the XRD and
XPS results confirm the stability of CdS nanorods. However, there is
obvious loss of Co 2p signal after recycling in lactic acid solution (Fig.
S7, SI†), and the depletion of Co species may result in the loss of
(
A.Q.Y., 39.6% at 400 nm) (Table S2, SI†).
To evaluate the reusability of CC4, recycling experiments were
photocatalytic activity. The loss of CoS
solution was further confirmed by the TEM images (Fig. S8a and b, SI†).
In the case of using Na S/Na SO as sacrificial agent, Co 2p signal
2
nanopartices in lactic acid
performed (Fig. S5a, SI†). When lactic acid is used as sacrificial agent,
2
2
3
2
Fig. 6. (a) UV–vis diffuse reflectance spectra (DRS), (b) the corresponding (αhv) versus hv curves, (c) photoluminescence spectra (excitation light: 350 nm), and (d)
transient photocurrent response in 0.5 M Na
2
SO
4
2
aqueous solution under visible light (λ > 400 nm) of pure CdS and CoS /CdS composites.
7

J. Tang, et al.
Applied Catalysis A, General 588 (2019) 117281
Fig. 7. (a) Extent of photocatalytic H
2
evolution, and (b) photocatalytic H
2
evolution rate over the as-prepared samples under visible light irradiation. Reaction
conditions: 100 mL (10 vol% LA) aqueous solution, 20 mg catalyst, 300 W Xe-lamp (λ ≥ 400 nm) for 3 h.
for visible-light-driven water splitting.
. Conclusion
4
In summary, CoS
surface of CdS nanorods using a simple in-situ solvothermal method.
With the deposition of an optimal amount CoS nanoparticles, CdS
nanorods exhibit high apparent quantum yield (39.6%, λ = 400 nm)
2
nanoparticles were successfully loaded on the
2
−
1
−1
and outstanding efficiency for H
2
production (˜58.1 mmol·g
h
)
under visible light, much higher than that of CdS nanorods deposited
−
1
−1
with 1 wt%Pt (19.8 mmol·g
h
). Based on the results of DFT cal-
culation and characterization, it is deduced that with class nature of
precious metals, cobalt disulfide acts as an effective co-catalyst that can
not only efficiently enhance the separation and migration of photo-
generated charge carriers but also offer ample surface sites for proton
reduction. This article demonstrates that CoS
ternative to noble metals as co-catalyst for the enhancement of photo-
catalytic H production under visible light irradiation. Moreover, the
2
is a commendable al-
Fig. 8. Possible mechanism of visible-light-driven H
2
production over CdS na-
norods modified with CoS nanoparticles using hole scavenger.
2
2
findings provide a new synthetic route for the coupling of zero di-
mensional nanoparticles with two dimensional semiconductors.
similar to that of fresh CC4 can still be observed after use (Fig. S7, SI†).
Also, uniform dispersion of CoS nanoparticles on CdS can be found in
the TEM image of CC4 after 4 cycles of use (Fig. S8c and d, SI†). The
2
Acknowledgements
2 2 3
results indicate the stability of CC4 when Na S/Na SO is adopted as
sacrificial agent.
This project was financially supported by the NSFC
Grants21725602, 21776064, and21671062), and Innovative Research
Groups of Hunan Province (Grant 2019JJ10001). C. T. Au thanks HNU
for an adjunct professorship.
Overall, CdS nanorods exhibit low activity in photocatalytic hy-
drogen production due to poor separation efficiency of photogenerated
charge carriers. When CoS
modified CdS nanorods exhibit higher hydrogen production efficiency
under visible light irradiation. It is because CoS nanoparticles can
(
2
nanoparticles are used as co-catalyst, the
2
Appendix A. Supplementary data
strengthen the absorption of visible light and enhance the separation
efficiency of photogenerated charge carriers. Furthermore, the loading
of CoS nanoparticles with class nature of precious metals can enrich
2
Supplementary material related to this article can be found, in the
online version, at doi:https://doi.org/10.1016/j.apcata.2019.117281.
the CdS surface with sites for proton reduction. Based on the above
results, a possible mechanism is proposed for HER under visible light as
illustrated in Fig. 8. Upon visible light illumination, electrons and holes
are generated on the surface of CdS nanorods. Then the photoexcited
charge carriers are subject to immediate separation and/or transition
under the influence of internal electric field. Part of the photogenerated
electrons in the CdS valence band (VB) transfer to the CdS conduction
band (CB) and ulteriorly migrate to the catalyst surface, leaving holes in
References
[
[
[
[
[
1] M.F. Kuehnel, D.W. Wakerley, K.L. Orchard, E. Reisner, Angew. Chem. Int. Ed. 54
(2015) 9627–9631.
2] K. Maeda, K. Teramura, D. Lu, T. Takata, N. Saito, Y. Inoue, K. Domen, Nature 440
(2006) 295–295.
3] X.F. Qian, D. Yue, Z. Tian, M. Reng, Y. Zhu, M. Kan, T. Zhang, Y. Zhao, Appl. Catal.
B Environ. 193 (2016) 16–21.
4] R.E. Blankenship, D.M. Tiede, J. Barber, G.W. Brudvig, G. Fleming, M. Ghirardi,
Science 332 (2011) 805–809.
the VB of CdS. Finally, lactic acid molecules (or Na
oxidized by the VB holes and water molecules are reduced to hydrogen
by the CB electrons. The presence of CoS nanoparticles on CdS can
lower the migration resistance of photogenerated electrons for facile
transit to CoS nanoparticles, consequently enhancing the separation of
charge carriers and promoting the photocatalytic performance of CdS
2 2 3
S/Na SO ) are
5] S.S. Yi, X.B. Zhang, B.R. Wulan, J.M. Yan, Q. Jiang, Energy Environ. Sci. 11 (2018)
2
3128–3156.
[6] X. Zou, Y. Zhang, Chem. Soc. Rev. 44 (2015) 5148–5180.
[7] J. Tian, Q. Liu, A.M. Asiri, X. Sun, J. Am. Chem. Soc. 136 (2014) 7587–7590.
[8] X.B. Chen, L. Liu, P.Y. Yu, S.S. Mao, Science 331 (2011) 746–750.
2
[9] Y. Zhu, T. Cao, C. Cao, J. Luo, W. Chen, L. Zheng, J. Dong, J. Zhang, Y. Han, Z. Li,
8

J. Tang, et al.
Applied Catalysis A, General 588 (2019) 117281
C. Chen, Q. Peng, D. Wang, Y. Li, ACS Catal. 8 (2018) 10004–10011.
10] Y. Shi, Y. Zhou, D.R. Yang, W.X. Xu, C. Wang, F.B. Wang, J.J. Xu, X.H. Xia,
[31] J. He, L. Chen, F. Wang, Y. Liu, P. Chen, C.T. Au, S.F. Yin, ChemSusChem 9 (2016)
624–630.
[32] J. He, L. Chen, Z.Q. Yi, C.T. Au, S.F. Yin, Ind. Eng. Chem. Res. 55 (2016)
8327–8333.
[33] J. He, L. Chen, Z.Q. Yi, D. Ding, C.T. Au, S.-F. Yin, Catal. Commun. 99 (2017)
79–82.
[34] Z. Wang, J. Peng, X. Feng, Z. Ding, Z. Li, Catal. Sci. Technol. 7 (2017) 2524–2530.
[35] H. Li, P. Deng, Y. Hou, Mater. Lett. 229 (2018) 217–220.
[36] H. Wang, Z. jin, Sustain. Energy Fuels 3 (2019) 173–183.
[37] D.A. Reddy, E.H. Kim, M. Gopannagari, R. Ma, P. Bhavani, D.P. Kumar, T.K. Kim,
ACS Sustain. Chem. Eng. 6 (2018) 12835–12844.
[
[
[
[
[
[
H.Y. Chen, J. Am. Chem. Soc. 139 (2017) 15479–15485.
11] S.P. Adhikari, Z.D. Hood, H. Wang, R. Peng, A. Krall, H. Li, V.W. Chen, K.L. More,
Z. Wu, S. Geyer, A. Lachgar, Appl. Catal. B Environ. 217 (2017) 448–458.
12] J. Yu, J. Low, W. Xiao, P. Zhou, M. Jaroniec, J. Am. Chem. Soc. 136 (2014)
8
839–8842.
13] H. Chen, D. Jiang, Z. Sun, R.M. Irfan, L. Zhang, P. Du, Catal. Sci. Technol. 7 (2017)
515–1522.
14] K. He, J. Xie, Z.Q. Liu, N. Li, X. Chen, J. Hu, X. Li, J. Mater. Chem. A 6 (2018)
3110–13122.
15] S. Ma, Y. Deng, J. Xie, K. He, W. Liu, X. Chen, X. Li, Appl. Catal. B Environ. 227
2018) 218–228.
1
1
[38] Y.Y. Fu, C. Zhu, C. Liu, M. Zhang, H. Wang, W. Shi, H. Huang, Y. Liu, Z. Kang, Appl.
Catal. B Environ. 226 (2018) 295–302.
(
[16] J.H. Yang, D. Wang, H.X. Han, C. Li, Acc. Chem. Res. 46 (2013) 1900–1909.
[39] Y. Zhu, L.F. Song, N. Song, M.X. Li, C. Wang, X.F. Lu, ACS Sustain. Chem. Eng. 7
[17] K. Wu, Z. Chen, H. Lv, H. Zhu, C.L. Hill, T. Lian, J. Am. Chem. Soc. 136 (2014)
(2019) 2899–2905.
7
708–7716.
18] L. Zhang, X. Fu, S. Meng, X. Jiang, J. Wang, S. Chen, J. Mater. Chem. A 3 (2015)
3732–23742.
[40] M.S. Faber, R. Dziedzic, M.A. Lukowski, N.S. Kaiser, Q. Ding, S. Jin, J. Am. Chem.
Soc. 136 (2014) 10053–10061.
[41] C.J. Chen, P.T. Chen, M. Basu, J. Mater. Chem. A 3 (2015) 23466–23476.
[42] D. Jiang, Z. Sun, H. Jia, D. Lu, P. Du, J. Mater. Chem. A 4 (2016) 675–683.
[43] F. Zhang, H.Q. Zhuang, J. Song, Y.L. Men, Y.X. Pan, S.H. Yu, Appl. Catal. B Environ.
226 (2018) 103–110.
[44] D. Lang, F. Cheng, Q. Xiang, Catal. Sci. Technol. 6 (2016) 6207–6216.
[45] S. Peng, L. Li, S.G. Mhaisalkar, M. Srinivasan, S. Ramakrishna, Q. Yan,
ChemSusChem 7 (2014) 2212–2220.
[
[
[
[
[
[
2
19] I. Majeed, U. Manzoor, F.K. Kanodarwala, M.A. Nadeem, E. Hussain, H. Ali,
A. Badshah, J.A. Stride, M.A. Nadeem, Catal. Sci. Technol. 8 (2018) 1183–1193.
20] J.M. Lee, H.B. Jin, I.Y. Kim, Y.K. Jo, J.W. Hwang, K.K. Wang, M.G. Kim, Y.R. Kim,
S.J. Hwang, Small 11 (2015) 5771–5780.
21] Z. Song, B. Hong, X. Zhu, F. Zhang, S. Li, J. Ding, X. Jiang, J. Bao, C. Gao, S. Sun,
Appl. Catal. B Environ. 238 (2018) 248–254.
22] M. Gopannagari, D.P. Kumar, D.A. Reddy, S. Hong, M.I. Song, T.K. Kim, J. Catal.
[46] S. Ji, T. Li, Z.D. Gao, Y.Y. Song, J.J. Xu, Chem. Commun. 54 (2018) 8765–8768.
[47] L. Shi, D. Li, P. Yao, J. Yu, C. Li, B. Yang, C. Zhu, J. Xu, Small 14 (2018) 1802716.
[48] Q. Pan, Y. Liu, L. Zhao, Chem. Eng. J. 351 (2018) 603–612.
[49] C. Yang, X. Liang, X. Ou, Q. Zhang, H.S. Zheng, F. Zheng, J.H. Wang, K. Huang,
M. Liu, Adv. Funct. Mater. 29 (2019) 1807971.
3
51 (2017) 153–160.
23] J. Chen, X.J. Wu, L. Yin, B. Li, X. Hong, Z. Fan, B. Chen, C. Xue, H. Zhang, Angew.
Chem. Int. Ed. 54 (2015) 1210–1214.
[24] X. Zhang, Z. Lai, C. Tan, H. Zhang, Angew. Chem. Int. Ed. 55 (2016) 8816–8838.
25] J.H. Lin, Y.H. Tsao, M.H. Wu, T.M. Chou, Z.H. Lin, J.M. Wu, Nano Energy 31 (2017)
[
[50] M. Cai, J. Han, Y. Lin, W. Liu, X. Luo, H. Zhang, M. Zhong, Electrochim. Acta 287
5
75–581.
26] C. Li, H. Wang, S.B. Naghadeh, J.Z. Zhang, P. Fang, Appl. Catal. B Environ. 227
2018) 229–239.
(2018) 1–9.
[
[51] Y. Xu, M.A.A. Schoonen, Am. Mineral. 85 (2000) 543–556.
[52] X.L. Yin, G.Y. He, B. Sun, W.J. Jiang, D.J. Xue, A.D. Xia, L.J. Wan, J.S. Hu, Nano
Energy 28 (2016) 319–329.
(
[
[
27] Y. Zhang, Z. Peng, S. Guan, X. Fu, Appl. Catal. B Environ. 224 (2018) 1000–1008.
28] B.C. Qiu, Q.H. Zhu, M.M. Du, L.G. Fan, M.Y. Xing, J.L. Zhang, Angew. Chem. Int.
Ed. 56 (2017) 1–6.
[53] Y. Pan, X. Cheng, L. Gong, L. Shi, T. Zhou, Y. Deng, H. Zhang, ACS Appl. Mater.
Interfaces 10 (2018) 31441–31451.
[
29] P. Tan, Y. Liu, A. Zhu, W. Zeng, H. Cui, J. Pan, ACS Sustain. Chem. Eng. 6 (2018)
[54] Y. Pan, X. Cheng, L. Gong, L. Shi, H. Zhang, Nanoscale 10 (2018) 20813–20820.
[55] C.J. Chen, P.T. Chen, Mrinmoyee Basu, J. Mater. Chem. A 3 (2015) 23466–23476.
[56] P. Garg, S. Kumar, I. Choudhuri, A. Mahata, B. Pathak, J. Phys. Chem. C 120 (2016)
7052–7060.
1
0385–10394.
[
30] Q. Yuan, D. Liu, N. Zhang, W. Ye, H. Ju, L. Shi, R. Long, J. Zhu, Y. Xiong, Angew.
Chem. Int. Ed. 56 (2017) 4206–4210.
9