Interstitial P-Doped CdS with Long-Lived Photogenerated Electrons for Photocatalytic Water Splitting without Sacrificial Agents

Article abstract of DOI:10.1002/adma.201705941

Photocatalytic hydrogen evolution from pure water is successfully realized by using interstitial P-doped CdS with rich S vacancies (CdS-P) as the photocatalyst in the absence of any electron sacrificial agents. Through interstitial P doping, the impurity level of S vacancies is located near the Fermi level and becomes an effective electron trap level in CdS-P, which can change dynamic properties of photogenerated electrons and thus prolong their lifetimes. The long-lived photogenerated electrons are able to reach the surface active sites to initiate an efficient photocatalytic redox reaction. Moreover, the photocatalytic activity of CdS-P can be further improved through the loading of CoP as a cocatalyst.

Full text of DOI:10.1002/adma.201705941

COMMUNICATION
Photocatalysis
www.advmat.de
Interstitial P-Doped CdS with Long-Lived Photogenerated
Electrons for Photocatalytic Water Splitting without
Sacrificial Agents
Rui Shi, Hui-Fang Ye, Fei Liang, Zhuan Wang, Kai Li, Yuxiang Weng, Zheshuai Lin,
Wen-Fu Fu, Chi-Ming Che, and Yong Chen*
position.[1] However, enormous studies
have demonstrated that CdS suffers from
rapid recombination of photogenerated
carriers and severe photocorrosion, which
seriously hinders the improvement of its
photocatalytic efficiency.[2] To solve these
problems, electron sacrificial agents have
been frequently added into the reaction
system to consume the photogenerated
holes, thereby increasing the survival
time of photogenerated electrons and
inhibiting the photocorrosion. However,
it should be emphasized that H2 produc-
tion in the presence of electron sacrificial
Photocatalytic hydrogen evolution from pure water is successfully realized by
using interstitial P-doped CdS with rich S vacancies (CdS-P) as the photocatalyst
in the absence of any electron sacrificial agents. Through interstitial P doping,
the impurity level of S vacancies is located near the Fermi level and becomes
an effective electron trap level in CdS-P, which can change dynamic properties
of photogenerated electrons and thus prolong their lifetimes. The long-lived
photogenerated electrons are able to reach the surface active sites to initiate an
efficient photocatalytic redox reaction. Moreover, the photocatalytic activity of
CdS-P can be further improved through the loading of CoP as a cocatalyst.
agents can only be considered as a “partial solar fuel”. Under
this condition, a portion of the solar energy has been stored
in the chemical reaction between electron sacrificial agents
and photogenerated holes.[3] Hydrogen evolution occurs at the
cost of the consumption of electron sacrificial agents. There-
fore, there is an urgent need to explore alternative ways to
achieve photocatalytic water splitting in the absence of electron
sacrificial agents.
Solar water splitting using semiconductor photocatalysts has
attracted increasing attention during the past several decades.
The development of visible-light responsive photocatalysts is
indispensable for utilizing sunlight efficiently. In this context,
CdS has been a hotspot for photocatalytic hydrogen production
because of its narrow bandgap and appropriate band structure
In principle, the presence of electron trap centers could
inhibit photogenerated electron–hole recombination and
increase the lifetime of photogenerated electrons. In most
cases, CdS exhibits n-type semiconductor characteristics due
to its intrinsic S vacancies. The impurity level of S vacancies
is located between the conduction band (CB) and Fermi level
because of the low concentration of S vacancies.[4] According
to the classical theory of semiconductor physics,[5] the impurity
level acts as a recombination center when it is positioned at a
deep level near the center of the band gap (Et ≈ Ei) or becomes
an electron trap center when it is located near the Fermi level
(Et ≈ EF). Based on Fermi distribution function 1
Dr. R. Shi, H.-F. Ye, Prof. W.-F. Fu, Prof. Y. Chen
Key Laboratory of Photochemical Conversion and Optoelectronic
Materials and HKU-CAS Joint Laboratory on New Materials
Technical Institute of Physics and Chemistry & University of Chinese
Academy of Sciences
Chinese Academy of Sciences
Beijing 100190, P. R. China
E-mail: chenyong@mail.ipc.ac.cn
F. Liang, Prof. Z. Lin
Key Laboratory of Functional Crystals and Laser Technology
Technical Institute of Physics and Chemistry & University of Chinese
Academy of Sciences
Chinese Academy of Sciences
Beijing 100190, P. R. China
Dr. Z. Wang, Prof. Y. Weng
1
f (E) =
(1)
Key Laboratory of Soft Matter Physics
Institute of Physics & University of Chinese Academy of Sciences
Chinese Academy of Sciences
E −EF
kT
e
+ 1
Beijing 100190, P. R. China
where f(E) is the probability of an electron occupied energy
level E, EF is the Fermi level and k is the Boltzmann constant,
the impurity levels below the Fermi level are almost filled with
electrons and cannot play the role of an electron trap center,
whilst those above the Fermi level are almost empty and suit-
able as electron trap centers. However, the higher the impu-
rity level is, the greater the probability that trapped electrons
can be re-excited to the conduction band through thermal
Dr. K. Li, Prof. C.-M. Che
State Key Laboratory of Synthetic Chemistry & Department of Chemistry
The University of Hong Kong
Pokfulam Road, Hong Kong SAR, P. R. China
The ORCID identification number(s) for the author(s) of this article
can be found under https://doi.org/10.1002/adma.201705941.
DOI: 10.1002/adma.201705941
©
1705941 (1 of 6)
2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Mater. 2018, 1705941

www.advancedsciencenews.com
www.advmat.de
excitation. Therefore, only when the impurity level is slightly
higher or equal to Fermi level, a substantial electron trap effect
can be expected. For these reasons, the S vacancies in CdS are
not effective electron trap centers in most cases.[6] Pristine CdS
exhibits poor photocatalytic activity for water-splitting due to
the lack of electrons trap center.[7] Therefore, it is necessary
to endow more S vacancies and promote the Fermi level to be
close to the level of the S vacancies. The position of Fermi level
of an n-type semiconductor in a strong ionization region can be
calculated using Equation (2)[8]
electron sacrificial agents to realize H2 evolution from water.
Herein, we report a new strategy to fabricate CdS with more
S vacancies using hydrazine hydrate as a reducing agent.
Through an interstitial P-doping treatment, the impurity level
of S vacancies was tuned to be in close proximity of the Fermi
level and became an effective electron trap, which led to the
generation of long-lived photogenerated electrons and resulted
in marked photocatalytic water splitting activity without using
electron sacrificial agents.
CdS was synthesized by reacting Cd(OAc)2 and thiourea in
the presence of hydrazine monohydrate. The addition of hydra-
zine monohydrate was crucial to the generation of S vacancies
in CdS. Hydrazine monohydrate easily decomposed to produce
H2, N2, and NH3 under high temperature and pressure reac-
tion conditions. Cd2+ of CdS crystals was slightly reduced by in
situ generated H2 during the hydrothermal process. As a result,
S vacancies were generated in order to balance the charge.[11]
The intense visible absorption of CdS with an absorption edge
at 580 nm, which is significantly redshifted from 515 nm for
CdS without defects, implicates the presence of rich S vacan-
cies (Figure S1, Supporting Information). CdS-P was then pre-
pared via a thermal phosphorization reaction with NaH2PO2
as the P source. As illustrated in Figure S2 in the Supporting
Information, X-ray diffraction (XRD) pattern of the obtained
CdS-P closely resembles that of pristine CdS. However, a
close inspection of the main diffraction peaks reveals that the
peaks position of CdS-P shift slightly toward a lower 2θ value,
which means an increase in the lattice parameters (Figure 1a).
The lattice distances of CdS-P are 0.360 and 0.339 nm
for {100} and {002} facets, respectively, in comparison to the
EF = Ei + kT ln(ND /ni )
(2)
where Ei is approximated as the intrinsic Fermi level (located
at the center of band gap), k is the Boltzmann constant, ND is
the donor impurity concentration and ni is the intrinsic carrier
concentration (a constant at the specified temperature). Based
on this equation, the Fermi level of CdS could be promoted
through doping with a donor impurity.
Elemental P, as an amphoteric impurity, has been doped into
semiconductors to form p–n junctions in the field of light-emit-
ting diodes because P doping can change the semiconductor type
and carrier concentration.[9] In addition, elemental P has also been
reported to play an important role in improving the performance
of widely studied photocatalysts such as TiO2 and g-C3N4.[10]
In contrast, this strategy has been seldom applied to CdS
to boost its photocatalytic activity. Very recently, Huang et al.
reported that gradient P-doping into CdS can generate an ori-
ented built-in electric-field for efficient extraction of carriers.[7]
In this system, however, Na2S and Na2SO3 are still required as
Figure 1. a) XRD patterns of CdS and CdS-P in the range of 24–29°. b) HRTEM and fast Fourier transform (FFT) images of CdS. c) HRTEM and FFT
images of CdS-P. d) HRTEM image of 5%CoP/CdS-P. e) STEM and f) elemental mapping images of 5%CoP/CdS-P.
©
Adv. Mater. 2018, 1705941
1705941 (2 of 6)
2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.advmat.de
corresponding values of 0.356 and 0.335 nm, respectively, for
CdS. For the preparation of CoP/CdS-P, Co(OH)2/CdS was
used as the precursor for the phosphorization treatment. The
phosphorization process led to not only successful doping of P
into CdS but also the transformation of Co(OH)2 to CoP. The in
situ loading of CoP onto the surface of CdS-P enabled the forma-
tion of an integrated Schottky junction and facilitated the separa-
tion and transfer of the photogenerated carriers.
Although transmission electron microscopy (TEM) images
display that the phosphorization process has no obvious effect
on the morphology and size of CdS (Figure S3, Supporting
Information), high-resolution TEM (HRTEM) images reveal
that the lattice distance of CdS-P is slighter larger than that
of undoped CdS (Figure 1b,c). The atomic ratio of Cd:S:P was
estimated to be 1:0.986:0.068 by energy dispersive X-ray spec-
troscopy analysis. A similar value for the amount of P was also
obtained by inductively coupled plasma atomic emission spec-
trometry (ICP-OE). For 5%CoP/CdS-P, it is obvious that the
CoP and CdS-P nanoparticles are in close contact with each
other (Figure 1d). The lattice distances of 0.340 and 0.189 nm
observed in the HRTEM image can be assigned to the diffraction
planes of CdS-P and CoP, respectively. Scanning TEM (STEM)
image and corresponding elemental mapping of 5% CoP/
CdS-P, as depicted in Figure 1e,f, clearly indicate the existence
of Cd, S, P, and Co elements in the sample and the uniform
distributions of both P and Co.
Furthermore, Raman spectra and X-ray photoelectron spectros-
copy (XPS) experiments were conducted to identify the doping
form of P in CdS. The Raman spectra of CdS, CdS-P, and 5%
CoP/CdS-P are depicted in Figure S6 in the Supporting Informa-
tion. The peaks at ≈300 and 600 cm−1 correspond to the first- and
second-order longitudinal optical (LO) modes, respectively.[12]
The Raman bands of CdS-P and 5%CoP/CdS-P are slightly
redshifted with respect to CdS, implying that P has been suc-
cessfully incorporated into the lattice rather than onto the
surface of CdS. It has been reported that the peak value of
the LO mode is sensitive to the lattice strain along the c axis.[13]
Thelatticecontractions(Δc/cvalues)werecalculatedtobe−0.2388,
−0.2765, and −0.2985 for CdS, CdS-P, and 5%CoP/CdS-P,
respectively (the calculation details are shown in the supporting
information). The increase in the lattice distance from CdS-P to
5%CoP/CdS-P is likely due to the formation of a Schottky junc-
tion between CdS and CoP through tight contact. The surface
chemical states of the samples were further
suspended in acetone are shown in Figure 2a. The emission
peaks exhibit no obvious shift after interstitial P doping. Two
main emission bands are observed at ≈518 and 770 nm. With
reference to the literature,[4,15] the band at 518 nm is attrib-
uted to a band to band transition; the band at 770 nm can be
assigned to deep level interstitial-site and antisite defects. The
transitions of emission levels are illustrated in Figure S8 in
the Supporting Information. Figure 2b displays time-resolved
photoluminescence (TRPL) spectra of CdS, CdS-P, and 5%CoP/
CdS-P suspended in acetone and monitored at 518 nm. The
plot on a semilog scale shows that the dynamics follow second-
order exponential decay behavior (Table S1, the Supporting
Information). Li and co-workers have reported that the emis-
sion processes are closely related to the trap level under a weak
excitation condition.[16] Upon excitation, the photogenerated
carriers are first captured by the trap level. Then, the trapped
charge carriers can be thermally re-excited to the conduction
and valence bands (VBs), reaching a dynamic equilibrium.
Subsequently, the carriers undergo two competition processes:
(i) photocatalytic reaction and (ii) radiative and nonradiative
recombination. As shown in Figure 2b, CdS-P has a threefold
longer carrier lifetime (4.58 ns) compared to CdS (1.53 ns).
The long-lived photogenerated carriers are due to the presence
of a more effective trap level. Since CdS-P is an n-type semi-
conductor, the S vacancies constitute the majority of defects.
Therefore, the S vacancies of CdS-P can be used as a better trap
level than that of CdS. Namely, when the electrons of CdS-P are
excited to its conduction band, the S vacancies beneath the con-
duction band can serve as electron trap level to capture most of
the electrons. As a result, the decay rate of electrons becomes
slower, and the lifetime is prolonged. In addition, the loading of
CoP decreases the carrier lifetime to 2.12 ns for 5%CoP/CdS-P
in comparison with CdS-P. This effect is attributed to an addi-
tional decay channel that is opened through electron transfer
from CdS-P to CoP.[17] Hence, the Schottky junction between
CdS-P and CoP facilitated the separation and transfer of the
carriers.
To ascertain the role of S vacancies as an effective trap level
for electrons after P doping, the relevant energy levels of the
electronic band were determined. The UV–visible diffuse reflec-
tance spectroscopy results are shown in Figure S1 in the Sup-
porting Information. From the absorption edges, the bandgap
of CdS-P is estimated to be smaller than that of CdS by 0.15
investigated by XPS (Figure S7, Supporting
Information). The P 2p signals of CdS-P are
at 130.1 and 130.9 eV for P 2p3/2 and P 2p1/2
,
respectively, indicating that P atoms exist in
the zero-valence state.[14] Additionally, the
formation of Cd-P bonds can be excluded
given that the binding energy of Cd is
almost unchanged. Collectively, it is reason-
able to infer that P atoms were interstitially
doped into the framework of CdS.
Photoluminescence (PL) spectra of the
obtained materials were subsequently
measured to examine the recombination of
photogenerated electron–hole pairs. Room
temperature PL spectra of CdS and CdS-P
Figure 2. a) Room-temperature PL spectra of CdS and CdS-P in acetone (excitation at 450 nm).
b) Room-temperature TRPL decay curves of CdS, CdS-P, and 5%CoP/CdS-P in acetone (excita-
tion at 450 nm).
©
1705941 (3 of 6)
2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Mater. 2018, 1705941

www.advancedsciencenews.com
www.advmat.de
eV. Moreover, the former clearly exhibits an absorption tail in
580–800 nm. Because the composition of the CdS CB remains
unchanged after P doping, which is further supported by density
functional theory (DFT) calculation (vide infra), the CB energy
level of CdS-P does not shift. Hence, we can deduce that the VB
position of CdS-P is 0.15 eV higher than that of CdS. Further-
more, we employed time-resolved transient mid-IR absorption–
excitation energy scanning spectra (TIRA–ESS)[18] technique to
determine the midgap impurity levels of CdS and CdS-P. Figure
S9 in the Supporting Information shows the TIRA–ESS for CdS
and CdS-P with a delay time of 1.5 ns after the excitation pulse.
The IR absorption intensity of CdS monotonically decreases with
an increase in the excitation wavelength and finally vanishes at
578 nm. Given the band gap of hexagonal phase CdS of 2.4 eV,
the absorption between 515 and 578 nm should arise from the
transition from its VB to the impurity level lying below the CB,
which is most likely the S vacancies level. Based on the 578 nm
transition, the gap between the S vacancies level and the CB
is estimated to be 0.25 eV. For CdS-P, the IR absorption inten-
sity monotonically decreased until ≈810 nm, which means that
there are electron occupied energy levels above the VB. It is pre-
sumed that these energy levels are caused by P doping, which is
distributed above the VB. This highest electron occupied state
is determined to be 1.79 eV below the CB. To further corrobo-
rate this, we performed UV photoelectron spectroscopy (UPS)
analysis of the VB spectra for both CdS and CdS-P. As shown in
Figure S10 in the Supporting Information, CdS-P demonstrated
an excess electron density compared with that of CdS, which
is attributed to a hybridized state in the bandgap induced by
P doping. The flatband potential (Vfb) can be quantified by the
Mott–Schottky (M–S) equation and is the difference between
Fermi level and water-reduction potential.[7] The M–S plots of
CdS and CdS-P are depicted in Figure S11 in the Supporting
Information. The Vfb of CdS is determined to be −0.79 eV,
which means that CdS exhibits n-type semiconductor characteris-
tics. Compared with CdS, the Vfb of CdS-P is upgraded to −0.94 eV,
proving the formation of a stronger n-type semiconductor, and
the Fermi level of CdS-P is almost coincident with the level
of S vacancies. On the basis of the above results, the schematic
energy levels are summarized in Figure 3.
To gain insight into the intrinsic reason, DFT calculations
were carried out to simulate both the electronic and geometric
structures (the details are shown in the Experimental Section
in the Supporting Information). In a unit cell of hexagonal
CdS, there are two tetrahedral voids and two octahedral voids.
Considering the larger atomic radius of P atom and principle
of structural stability, one P atom is interstitially doped into
the octahedral voids in a 2 × 2 × 2 supercell. So the atomic
ratio of Cd:S:P is 1:1:0.06 in calculation model. Based on the
Eform equation, the simulative energies of the CdS supercell,
CdS-P supercell and isolated P atom are −25080.29, −25259.62,
and −176.47 eV, respectively. The calculated Eform of −2.86 eV
implies an exothermic reaction. Therefore, the interstitial P
doping of CdS is stable. The calculated lattice parameters a
and c for both CdS-P and CdS revealed a slight lattice expan-
sion upon P doping. (Table S2 and Figure S12, Supporting
Information). According to the experimental results and DFT
calculations, P atoms are interstitially doped into the octahedral
voids of the CdS cell and slightly change the intrinsic crystal
Figure 3. Schematic energy levels and variations in the Fermi levels of
CdS and CdS-P.
parameters. The variations in the band structures of CdS and
CdS-P are shown in Figure S13 in the Supporting Information.
The zero energy corresponds to the highest electron occupied
state. For the CdS supercell, the band gap is estimated to be
1.256 eV. The band gap of CdS-P is not distinct because of the
appearance of the electron level state in the band gap. To further
clarify the effect of P doping on the electronic structure, the par-
tial density of states (PDOS) for the s, p, d orbitals are shown in
Figure S14 in the Supporting Information. For the CdS super-
cell, the CB is mainly made up of Cd 5s and S 3p orbitals, while
the VB is mainly composed of S 3p orbitals. On the other hand,
for CdS-P supercell, the compositions of the CB are almost
unchanged, while the VB constitutes not only S 3p orbital but
also a small amount of P 3p orbital. Importantly, P doping
induces an electron occupied state in the band gap, which com-
prises hybridized Cd 5s, S 3p, and P 3p orbitals. Based on the
position of the zero energy, most of these hybridized states are
electron occupied states, which is consistent with TIRA–ESS
and UPS VB results. Due to the existence of the electron occu-
pied state in the band gap, the IR absorption in the TIRA–ESS
can be extended to 810 nm. According to the results of the DFT
calculation and experimental data, the schematic energy levels
and variation in the Fermi level are illustrated in Figure 3.
After P doping, the hybridized state appears in the band gap
and the Fermi Level is raised to approach the level of S vacan-
cies. Hence, the S vacancies of CdS-P become excellent electron
trap centers. As a result, the decay rate of the photogenerated
electrons is slower, and the lifetime is longer.
The obtained CdS, CdS-P, and X%CoP/CdS-P samples were
tested for pure water splitting under visible-light irradiation
without using electron sacrificial agents. As shown in Figure 4a,
pristine CdS does not show H2 production activity, while CdS-P
has moderate H2 production activity (76 µmol hg−1). The photo-
generated carriers in CdS-P with prolonged lifetime arising
from suppressed electron–hole recombination are conceived to
reach the surface active sites to initiate an efficient photocata-
lytic redox reaction.[19] When a certain amount of cocatalyst CoP
was loaded onto CdS-P, the H2 evolution activity was improved
significantly, and 5%CoP/CdS-P achieved the maximum
©
Adv. Mater. 2018, 1705941
1705941 (4 of 6)
2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.advmat.de
Figure 4. a) Photocatalytic H2 evolution rates over various samples under an LED. b) EPR responses of the DMPO–OH spin adduct in a 5%CoP/
CdS-P suspension. c) Time dependence of the absorption intensity of H2O2 over the 5%CoP/CdS-P suspension during photocatalytic water splitting;
the content of H2O2 was measured by UV–vis spectrophotometry. In an acidic environment (pH = 1–2), Fe2+ can be oxidized into Fe3+ by H2O2, and
Fe3+ can be combined with KSCN to form [Fe(SCN)]2+. [Fe(SCN)]2+ has the maximum absorption peak at 475 nm. d) Photocatalytic H2 evolution rate
over 5%CoP/CdS-P, 5%CoP/CdS-P + HC (3 mg), and 5%CoP/CdS-P + PFDL(50 µL).
activity (231 µmol hg−1). This improvement is attributed to the
facilitated charge carrier separation in terms of the electron
transfer from CdS-P to CoP.
Attempts to analyze the oxidation products of water split-
ting using 5%CoP/CdS-P were also made. The characteristic
signals of ·OH in the suspension were observed through
in situ EPR spectroscopy (Figure 4b). This finding suggests
the pathway of single electron oxidation of water to produce
(HC) and perfluorodecalin (PFDL).[22] As shown in Figure 4d,
the addition of HC and PFDL increased the H2 production
activity by 42% and 157%, respectively. These findings imply
that O2 might be generated, but rapidly undergoes recombina-
tion with evolved H2. In addition, the stability test indicated
that after four cycles in pure water, the photocatalytic activity of
5%CoP/CdS-P maintained 65% of its initial activity (Figure S15,
Supporting Information). The quantity of H2 produced was
4.8 times the content of P, so elemental P was not consumed
as an electron sacrificial agent during the photocatalytic reac-
tion. XRD patterns and TEM images of the recovered photo-
catalysts are consistent with those of the as-prepared sample
(Figures S16 and S17, Supporting Information). However,
·
OH. It is noted that the oxidation potential for this half reac-
tion is 2.1 V, whereas the VB position of CdS is located at
1.2 V versus NHE at pH = 7,[20] implying that the oxidation
of water to OH by photogenerated holes of CdS is thermo-
·
dynamically not favored. However, recent research demon-
strated that the active sites of a catalyst have the ability to acti-
vate the target compound with lower energy demands.[20,21] In
the 5%CoP/CdS-P suspension system, it can be speculated
that by means of surface activation, water molecules could
be oxidized into OH through a single electron pathway by
photogenerated holes. With the photocatalytic reaction pro-
gressing, a small amount of H2O2 was further observed,
and the concentration of H2O2 gradually increased over 14 h
(Figure 4c). The formation of H2O2 is tentatively attributed to
the combination of two OH. In contrast, our analysis shows
that no O2 evolved in this reaction system. To verify whether
there was a reverse reaction between evolved H2 and O2
to produce H2O, water splitting reactions were performed in
the presence of the oxygen transfer reagents, hemin chloride
2−
a trace amount of SO4 was detected after the fourth cycle,
likely due to oxidization of the photocatalyst by some of the
photogenerated holes. It is surmised that this photocorrosion
caused the drop in activity.
In conclusion, we have demonstrated that interstitial
P-doped CdS with rich S vacancies can be used for photocata-
lytic pure water splitting without using any electron sacrificial
agents. Long-lived photogenerated electrons are the essen-
tial reason for initiating photocatalytic water splitting. The
present study provides a pathway to construct efficient solar
water-splitting systems. Attempts to separate evolved H2 and
O2 effectively and to inhibit photocorrosion are the focus of
our next work for achieving a stable overall water splitting
reaction.
©
1705941 (5 of 6)
2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Mater. 2018, 1705941

www.advancedsciencenews.com
www.advmat.de
[9] a) X. Fang, J. H. Li, D. X. Zhao, D. Z. Shen, B. H. Li, X. H. Wang,
J. Phys. Chem. C 2009, 113, 21208; b) Y. W. Heo, Y. W. Kwow,
Y. Li, S. J. Pearton, D. P. Norton, Appl. Phys. Lett. 2004, 84, 3474;
c) A. Allenic, W. Guo, Y. B. Chen, M. B. Katz, G. Y. Zhao, Y. Che,
Z. D. Hu, B. Liu, S. B. Zhang, X. Q. Pan, Adv. Mater. 2007, 19, 3333;
d) W. J. Lee, J. Kang, J. Chang, Phys. Rev. B 2006, 73, 024117.
[10] a) W. Liu, L. L. Cao, W. R. Cheng, Y. J. Cao, X. K. Liu, W. Zhang,
X. L. Mou, L. L. Jin, X. S. Zheng, W. Che, Q. H. Liu, T. Yao,
S. Q. Wei, Angew. Chem., Int. Ed. 2017, 56, 9312; b) J. G. Yu,
L. Zhang, Z. Zheng, J. Zhao, Chem. Mater. 2003, 15, 2280;
c) N. O. Gopal, H. H. Lo, T. F. Ke, C. H. Lee, C. C. Chou, J. D. Wu,
S. C. Sheu, S. C. Ke, J. Phys. Chem. C 2012, 116, 16191; d) R. Zheng,
L. Lin, J. Xie, Y. Zhu, Y. Xie, J. Phys. Chem. C 2008, 112, 15502;
e) S. Hu, L Ma, J. You, F. Li, Z. Fan, F. Wang, D. Liu, J. Gui, RSC
Adv. 2014, 4, 21657; f) Z. F. Hu, Z. R. Shen, J. C. Yu, Green Chem.
2017, 19, 588.
Supporting Information
Supporting Information is available from the Wiley Online Library or
from the author.
Acknowledgements
The authors thank the financial support from the Strategic Priority
Research Program of the Chinese Academy of Sciences (XDB17000000)
and the Natural Science Foundation of China (21773275, 21371175).
Conflict of Interest
The authors declare no conflict of interest.
[11] X. Y. Zhang, Z. Zhao, W. W. Zhang, G. Q. Zhang, D. Qu, X. Miao,
S. R. Sun, Z. C. Sun, Small 2016, 12, 793.
[12] L. Zeiri, I. Patla, S. Acharya, Y. Golan, S. Efrima, J. Phys. Chem. C
2007, 111, 11843.
Keywords
photocatalysis, P-doped CdS, sacrificial-agents-free catalysts, water splitting
[13] C. Hu, X. H. Zeng, J. J. Cui, H. T. Chen, J. F. Lu, J. Phys. Chem. C
2013, 117, 20998.
Received: October 12, 2017
Revised: November 18, 2017
Published online:
[14] a) P. X. Qiu, C. M. Xu, N. Zhou, H. Chen, F. Jiang, Appl. Catal.,
B 2018, 221, 27; b) W. B. Li, Y. P. Zhang, G. D. Tian, S. Y. Xie,
Q. Q. Xu, L. Wang, J. M. Tian, Y. Y. Bu, J. Mol. Catal. A: Chem.
2016, 423, 356; c) W. B. Li, J. G. Yue, F. X. Hua, C. Feng, Y. Y. Bu,
Z. Y. Chen, Mater. Res. Bull. 2015, 70, 13.
[15] a) W. F. Liu, C. G. Jin, C. Jia, L. Z. Yao, W. L. Cai, X. G. Li, Chem.
Lett. 2004, 33, 228; b) G. Z. Shen, J. H. Cho, J. K. Yoo, G. C. Yi,
C. J. Lee, J. Phys. Chem. B 2005, 109, 9294.
[16] X. L. Wang, Z. C. Feng, J. Y. Shi, G. Q. Jia, S. Shen, J. Zhou, C. Li,
Phys. Chem. Chem. Phys. 2010, 12, 7083.
[1] a) Q. Li, B. D. Guo, J. G. Yu, J. R. Ran, B. H. Zhang, H. J. Yan,
J. R. Gong, J. Am. Chem. Soc. 2011, 133, 10878; b) K. F. Wu,
H. M. Zhu, Z. Liu, W. R. Cordoba, T. Q. Lian, J. Am. Chem. Soc.
2012, 134, 10337.
[2] a) A. Kudo, Y. Miseki, Chem. Soc. Rev. 2009, 38, 253; b) Y. Hu,
X. Cao, L. Yu, Y. Y. Wang, J. Ning, S. Xu, X. W. Lou, Angew. Chem.,
Int. Ed. 2013, 52, 5636.
[17] C. M. Li, Y. H. Du, D. P. Wang, S. M. Yin, W. G. Tu,
Z. Chen, M. Kraft, G. Chen, R. Xu, Adv. Funct. Mater. 2017, 27,
1604328.
[3] Y. Ma, X. L. Wang, Y. S. Jia, X. B. Chen, H. X. Han, C. Li, Chem. Rev.
2014, 114, 9987.
[4] X. L. Xu, Y. Y. Zhao, E. J. Sie, Y. H. Lu, S. A. Ekahana, X. Ju, Q. Jiang,
J. B. Wang, H. D. Sun, T. C. Sum, C. H. Huan, Y. P. Feng, Q. Xiong,
ACS Nano 2011, 5, 3660.
[18] a) Y. Mi, Y. X. Weng, Sci. Rep. 2015, 5, 11482; b) M. Zhu, Y. Mi,
G. B. Zhu, D. Y. Li, Y. P. Wang, Y. X. Weng, J. Phys. Chem. C 2013,
117, 18863.
[19] Z. Y. Zhang, J. D. Huang, Y. R. Fang, M. Y. Zhang, K. C. Liu, B. Dong,
Adv. Mater. 2017, 29, 1606688.
[5] E. K. Liu, B. K. Zhu, J. S. Luo, Semiconductor Physics, Electronic
Industry Press, Beijing, China 2008, p. 162.
[20] S. N. Habisreutinger, L. S. Mende, J. K. Stolarczyk, Angew. Chem.,
Int. Ed. 2013, 52, 7273.
[6] a) Q. Xiao, C. Xiao, Appl. Surf. Sci. 2009, 255, 7111; b) J. J. Ramsden,
S. E. Webber, M. Gratzel, J. Phys. Chem. 1985, 89, 2740; c) P. Zhang,
L. Gao, J. Colloid. Interface Sci. 2003, 266, 457.
[21] a) H. Li, J. Shang, Z. H. Ai, L. Z. Zhang, J. Am. Chem. Soc. 2015,
137, 6393; b) K. Zhao, L. Z. Zhang, J. J. Wang, Q. X. Li, W. W. He,
J. J. Yin, J. Am. Chem. Soc. 2013, 135, 15750.
[7] H. M. Huang, B. Y. Dai, W. Wang, C. H. Lu, J. H. Kou, Y. Ni,
L. Z. Wang, Z. Z. Xu, Nano Lett. 2017, 17, 3803.
[8] E. K. Liu, B. K. Zhu, J. S. Luo, Semiconductor Physics, Electronic
Industry Press, Beijing, China 2008, p. 81.
[22] a) Z. Li, B. Tian, W. L. Zhen, Y. Q. Wu, G. X. Lu, Appl. Catal., B
2017, 203, 408; b) M. Wang, Z. Li, Y. Q. Wu, J. T. Ma, G. X. Lu,
J. Catal. 2017, 353, 162; c) B. Tian, B. J. Yang, J. Li, Z. Li, W. L. Zhen,
Y. Q. Wu, G. X. Lu, J. Catal. 2017, 350, 189.
©
Adv. Mater. 2018, 1705941
1705941 (6 of 6)
2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim