Journal of Catalysis 266 (2009) 165–168

Contents lists available at ScienceDirect

Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat

Visible-light-driven hydrogen production with extremely high quantum

efficiency on Pt–PdS/CdS photocatalyst
Hongjian Yan a,1, Jinhui Yang a,b,1, Guijun Ma a,b, Guopeng Wu a, Xu Zong a,b, Zhibin Lei a, Jingying Shi a,
Can Li a,*
a
State Key Laboratory of Catalysis, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, 457 Zhongshan Road, Dalian 116023, China
b
Graduate University of Chinese Academy of Sciences, Beijing 100049, China

a r t i c l e i n f o a b s t r a c t

Article history: To efficiently convert solar energy into chemical energy by artificial photosynthesis, we need to develop
Received 3 May 2009 visible-light-responsive photocatalysts with a high quantum efficiency (QE). Here we report that an arti-
Revised 24 June 2009 ficial photocatalyst (Pt–PdS/CdS) can achieve a QE up to 93% in photocatalytic H2 production in the pres-
Accepted 29 June 2009
ence of sacrificial reagents under visible light irradiation, and is very stable under the photocatalytic
Available online 3 August 2009
reaction conditions. The extremely high QE could be achieved by loading as low as 0.30 wt% of Pt and
0.13 wt% of PdS as cocatalysts on CdS.
Keywords:
Ó 2009 Elsevier Inc. All rights reserved.
Photocatalysis
Hydrogen
Cocatalyst
Quantum efficiency
Cadmium sulfide

1. Introduction lytic activity even without cocatalysts, in most cases, cocatalysts

enable or increase the photocatalytic activity of a semiconductor
The photocatalytic production of hydrogen from water to store [2]. The cocatalysts could function as active sites for reduction or
solar energy, mimicking thus natural photosynthesis by converting oxidation and/or suppress the back reaction of water formation
light into chemical energy, has attracted increased attentions re- from evolved H2 and O2, etc. [2]. Noble metals (e.g., Pt, Pd, Ru,
cently [1–3]. Most of the developed photocatalysts capable of split- and Rh) [2], transition metal oxides (e.g., NiO [4] and Rh2yCryO3
ting water can only utilize ultraviolet irradiation constituting only [1]) and transition metal sulfides (e.g., MoS2 [8]) have been found
4% of the incoming solar energy [2–4]. And many researchers have to be effective for facilitating H2 evolution. Some transition metal
focused on the development of visible-light-responsive photocata- oxides (e.g., RuO2 and IrO2) are found to be good oxidation cocata-
lysts in order that the abundant visible light (k > 420 nm) which lysts for O2 evolution, although RuO2 is also found to be a reduction
occupies about 43% of the solar energy can be utilized [2,3,5]. A cocatalyst [9]. Up to date, only a few researches have shown that
key issue with regard to the conversion of solar energy into chem- suitable dual cocatalysts could increase the photocatalytic activity
ical energy is the QE of photocatalysts [3]. In the primary course of of a semiconductor catalyst [10].
natural photosynthesis, the overall QE can be as high as 95% [6]. Here we report that CdS co-loaded with Pt and PdS (denoted as
Among the reported visible-responsive chemical photocatalysts Pt–PdS/CdS) can achieve a QE up to 93% at 420 nm for photocata-
up to now [2,3], the highest QE for H2 production is still 60% over lytic H2 production using sulfide and sulfite as sacrificial reagents.
Pt-loaded CdS catalyst [7]. Furthermore, the Pt–PdS/CdS catalyst does not show deactivation
The previous researches on photocatalytic water splitting using for H2 evolution for longer than 100 h.
chemical photocatalysts have been focused on the development of
semiconductor-based photocatalytic materials, while the explora- 2. Experimental
tion of cocatalysts is less concerned with this issue [3]. Although
certain photocatalysts have been found to exhibit high photocata- 2.1. Preparation of the catalysts

All the reagents were of analytical grade, and were used with-
* Corresponding author. Fax: +86 411 84694447.
E-mail address: [email protected] (C. Li).
out further purification.
URL: http://www.canli.dicp.ac.cn (C. Li). The CdS nanoparticles were prepared as follows: an aqueous
1
Both authors contributed equally to this work. solution of Na2S (800 mL, 0.14 M) was added slowly to Cd(OAc)2

0021-9517/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2009.06.024
166 H. Yan et al. / Journal of Catalysis 266 (2009) 165–168

solution (1000 mL, 0.14 M) under vigorous stirring. The yellow The apparent quantum efficiency (QE) was measured under the
mixture was stirred for 24 h and kept for an additional 24 h. The same photocatalytic reaction condition with irradiation light at
resulting yellow slurry was filtered. The wet solid was suspended 420 nm by using combined band-pass (Kenko) and cut-off filters
in pure water (120 mL) and transferred to a Teflon-lined stainless (HOYA) and 300 W Xe lamp, and the QE was calculated according
steel autoclave (150 mL) and heated at 473 K for 72 h (hydrother- to Eq. (1):
mal treatment). The yellow solid was filtered, washed with water
number of reacted electrons
and ethanol subsequently, filtered again, and kept under vacuum QE ½% ¼  100
number of incident photons
at 368 K for 24 h.
The loading or co-loading of Pt and Pd on CdS producing Pt/CdS, number of evolved H2 molecules  2
¼  100 ð1Þ
Pd/CdS, and Pt–Pd/CdS was done by photodeposition. Thus, for number of incident photons
example, H2PtCl6 aqueous solution (0.0019 M) was added to a sus-
The number of incident photons was 3.54  1017 photons/s which
pension of CdS (1 g, as prepared above) dispersed in acetic acid
was measured by using a calibrated Si photodiode (HAMAMATSU
(AcOH, 80 mL, 0.25 M) and containing Cd(Ac)22H2O (0.5 g). The
S2281). The number of incident photons was also calibrated with
mixture was deaerated under vacuum and then irradiated with
ferrioxalate solution in the same device [11].
300 W Xe lamp for 30 min. The resulting suspension was then fil-
tered, washed with water, and dried at 363 K under vacuum.
3. Results and discussion
The loading of PdS on CdS producing PdS/CdS was realized by
in-situ deposition. Thus, for example, a PdCl2 aqueous solution
Well-crystallized cadmium sulfide (CdS) nanoparticles possess
(0.0011 M) was added dropwise to a suspension of CdS powder
hexagonal structures and diameters of approximately 40–60 nm
(0.3 g, as prepared above) dispersed in Na2S aqueous solution
(Supplementary material, Figs. 1 and 2) and their UV–vis diffuse
(0.5 M) just before the photocatalytic reaction.
reflectance spectrum shows an absorption edge at about 540 nm
The co-loading of Pt and PdS on CdS producing Pt–PdS/CdS was
(2.3 eV) (Supplementary material, Fig. 3).
realized similar to the loading of PdS except for the additional
In Table 1, the photocatalytic activities (or QE) for H2 evolution
addition of a H2PtCl6 aqueous solution (0.0019 M).
over CdS, Pt/CdS, Pd/CdS, Pt–Pd/CdS, and Pt–PdS/CdS catalysts are
given. CdS without the cocatalysts loaded exhibits very low photo-
2.2. Characterization catalytic activity for H2 evolution (ca. 0.023 mmol h1), and the
activity is enhanced markedly by loading noble metals such as
The phase of the as-prepared CdS was examined by X-ray dif- Pd and Pt as cocatalysts, while the activity is hardly increased fur-
fraction (Rigaku; MiniFlex diffractometer with a Cu Ka irradiation ther when Pt and Pd are co-loaded. When PdS is employed as the
source). Diffraction patterns were collected from 20° to 80° at a cocatalyst, the photocatalytic activity is even higher than that of
speed of 5° min1. The Ultraviolet-visible diffuse reflectance spec- Pt/CdS, achieving 280 times higher activity than that of CdS alone.
tra (UV–vis DRS) were obtained using a JASCO V-550 spectropho- Most interestingly, the activity is enhanced further to 380 times
tometer. The size and morphology were examined with Scanning when Pt and PdS are co-loaded on CdS, and the H2 evolution rate
Electron Microscopy (SEM) and Transmission Electron Microscopy is estimated to be as high as 8.77 mmol h1, corresponding to a
(TEM). SEM images were obtained on a Quanta 200F microscope QE of 93% at 420 nm. In our previous work, the co-loading of Pt
(FEI Company) with the accelerating voltage of 0.5–30 kV. TEM and PdS on CdS could result in an enhancement of activity in the
images were obtained on Tecnai G2 Spirit (FEI Company) micros- direct splitting of H2S to H2 and S in organic media, although the
copy with the accelerating voltage of 120 kV. QE is only about 30% [12].
Fig. 1 shows the reaction time courses for H2 evolution over Pt/
2.3. Photocatalytic reaction test CdS, PdS/CdS, and Pt–PdS/CdS photocatalysts. PdS/CdS and Pt–PdS/
CdS exhibit very stable activity during the 25-h irradiation. How-
The photocatalytic reactions were carried out in a Pyrex reac- ever, the rate of H2 production over Pt/CdS gradually decreases
tion cell connected to a closed gas circulation and evacuation sys-
tem. Photocatalyst, 0.3 g, was dispersed in 200 mL of aqueous
solution containing 0.5 M Na2S and 0.5 M Na2SO3 as the sacrificial
reagents. The suspension was then thoroughly degassed and irradi-
ated by a 300 W Xe lamp (ILC technology; CERAMAX LX-300)
which is equipped with an optical filter (HOYA, k > 420 nm) to
cut off the light in the ultraviolet region. The temperature of the
reaction solution was maintained at 12–18 °C by a flow of cooling
water. The amount of hydrogen evolved was determined with on-
line gas chromatography (Shimadazu; GC-8A, MS-5 A column, TCD,
Ar Carrier).

Table 1
Photocatalytic H2 evolution over CdS loaded with various cocatalysts.

Catalyst H2 evolution rate (mmol h1)a QE (%)

CdS 0.023 0.2
Pd/CdS 3.82 40
Pt/CdS 4.80 51
Pt–Pd/CdS 4.95 53
PdS/CdS 5.99 64
Pt–PdS/CdS 8.77 93
Fig. 1. The time courses of H2 evolution under visible irradiation over Pt/CdS (N),
a
Experiments were done using 0.3 g of photocatalysts powder suspended in PdS/CdS (), and Pt–PdS/CdS (j). Catalyst: 0.012 g; 0.5 M Na2S–0.5 M Na2SO3
200 mL of 0.5 M Na2S-0.5 M Na2SO3 solution in a pyrex glass cell under visible light aqueous solution (200 mL); Light source: Xe lamp (300 W) with an optical filter
irradiation (k > 420 nm). Light source: 300 W Xe lamp. Loading amount: 1 wt%. (k > 420 nm).
 H. Yan et al. / Journal of Catalysis 266 (2009) 165–168 167

and an evident drop in the activity is found after an irradiation for

10 h. This could be largely attributed to the photocorrosion of CdS
[7]. The lifetime experiments of the photocatalysts show that the
CdS and Pt/CdS are evidently deactivated, while the PdS/CdS and
Pt–PdS/CdS exhibit stable activity even after a photocatalytic reac-
tion for longer than 100 h.
Fig. 2 shows the dependence of the photocatalytic activity on
the loading amount of Pt and PdS. For Pt/CdS, the rate of H2 evolu-
tion is increased with the increase of Pt-loading from 0 to 0.65 wt%.
When the loading amount of Pt is increased to 0.65 wt%, the activ-
ity reaches a maximum, 4.80 mmol 1 with a QE of 51%, at 420 nm.
This activity is close to that recently reported for Pt/CdS by Domen
et al. [7]. The rate of H2 evolution over PdS/CdS shows a similar
trend, but only 0.13 wt% PdS achieved a maximum activity as high
as 5.99 mmol h1, corresponding to a QE of 64% at 420 nm. Surpris-
ingly, we found that the activity can be greatly increased further
when Pt is co-loaded with PdS on CdS. The maximum activity of
8.77 mmol h1 with a QE of 93% can be achieved by co-loading
only 0.30 wt% Pt and 0.13 wt% PdS on CdS. Fig. 3. X-ray photoelectron spectra (XPS) of Pd 3d in the Pt–PdS/CdS catalyst
collected after an irradiation with visible light (k > 420 nm) for 3 h in 0.5 M Na2SO3
The chemical states of Pd and Pt on catalysts after photocata-
(upper) and 0.5 M Na2S–0.5 M Na2SO3 (lower) aqueous solutions. The solid curves
lytic reaction were examined by X-ray photoemission spectros- are the smoothing original data. The dashed lines are composite fits of PdII and Pd0.
copy (XPS). The results show that the binding energy of Pt (4f7/ The lines were fitted by 80% Gaussian and 20% Lorentz functions after Shirley
2) is consistent with Pt0 (Supplementary material, Fig. 4) [13], background subtraction. Palladium 3d5/2/3d3/2 intensity ratio was set to 3:2 and the
and that the dominant valence for the palladium of PdS on CdS col- Binding Energy difference was 5.25 eV. The data were calibrated by C1s at 284.6 eV.
The atomic ratio of PdII to Pd0 is determined according to the peak area [15].
lected after the photocatalytic reaction is PdII, while a small portion
of metallic Pd also exists, indicating a partial photoreduction of PdS
The direction of charge transfer between the components could
during the photocatalytic reaction. The amount of Pd0 is only about
be predicted according to their energy band structures [16]. It has
1/20 of that of PdII in PdS (Fig. 3) [14].
been reported that PdS is an n-type semiconductor with a band gap
To testify whether such a low amount of metallic Pd is due to
of 1.60 eV [17]. The conduction band edge of PdS is estimated to be
the reaction of Na2S with Pd during the post-treatment, the PdS/
0.5 V vs. SCE (0.26 V vs. NHE) (Supplementary material, Fig. 5)
CdS material was collected after photocatalytic reaction. It was
by Mott-Schottky analysis [18], and the valence band potential of
redispersed in Na2SO3 aqueous solution (without Na2S) and col-
PdS deduced from the conduction band and the band gap
lected again after an irradiation for 3 h. The XPS spectra shown
(1.6 eV) is approximately +1.34 V vs. NHE, which is less positive
in Fig. 3 reveal the coexistence of metallic Pd and PdS with the ratio
than that of CdS (+1.5 V) [2]. This implies that the hole transfer
of Pd0 to PdII in PdS up to 2.8. This indicates the formation of metal-
from CdS to PdS is a favorable process, and that the PdS could act
lic Pd during the photocatalytic reaction. On the other hand, the
as an oxidation cocatalyst here. The conduction band edge of PdS
reduction potential of PdS/Pd (EPdS/Pd) (0.84 V vs. NHE), compara-
is less negative than the H2O/H2 redox potential (0.80 V vs.
ble to that of H2O/H2(0.80 V vs. NHE), makes the photoreduction
NHE) under the photocatalytic condition (pH = 13.6), so PdS by it-
of PdS possible. Thus, a QE as high as 64% in photocatalytic hydro-
self displays no photocatalytic activity for H2 production.
gen production achieved on PdS/CdS might be due to the syner-
Therefore, the simultaneous existence of PdS acting as an oxida-
getic effect of PdS and metallic Pd.
tion cocatalyst and Pt (or Pd) acting as a reduction cocatalyst is
supposed to be beneficial for the efficient separation and transfer
of the photoexcited electrons and holes, thus contributing to the
extremely high QE. The PdS acting as the oxidation cocatalyst
can also protect CdS from photocorrosion, and can make the PdS/
CdS and Pt–PdS/CdS very stable under the photocatalytic reaction
conditions. Further work is underway to establish the roles of the
dual cocatalysts in photocatalytic H2 production.

4. Conclusions

In summary, an extremely high QE for H2 production has been

achieved over Pt–PdS/CdS photocatalysts. The Pt–PdS/CdS catalyst
demonstrates the possibility of realizing visible-light-responsive
photocatalytic hydrogen production with a QE approaching the le-
vel of natural photosynthesis. The strategy to achieve high QE by
co-loading suitable dual cocatalysts, especially those functioning
as oxidation and reduction cocatalysts, respectively, will be of con-
siderable importance in the design and preparation of highly active
Fig. 2. The rate of H2 evolution as a function of the loading amount of Pt and PdS. x
denotes the loading amount. The lower curve shows the dependence of activity of photocatalysts for solar energy conversion.
Pt/CdS on the loading amount of Pt; the middle curve shows the dependence of
activity of PdS/CdS on the loading amount of PdS. The uppermost curve shows the Acknowledgments
dependence of the activity of Pt-0.13 wt% PdS/CdS on the loading amount of Pt. The
highest QE (93%) was obtained through co-loading 0.13 wt% PdS and 0.3 wt% Pt on
CdS. Catalyst: 0.3 g; 0.5 M Na2S–0.5 M Na2SO3 aqueous solution (200 mL); Light This work was finacially supported by National Program on Key
source: Xe lamp (300 W) with an optical filter (k > 420 nm). Basic Research Project (Grant No.: 2009CB220010) and Solar En-
168 H. Yan et al. / Journal of Catalysis 266 (2009) 165–168

ergy Action Project of Chinese Academy of Sciences. The authors [6] V. Sundström, T. Pullerits, R. van Grondelle, J. Phys. Chem. B 103 (1999) 2327.
[7] N.Z. Bao, L.M. Shen, T. Takata, K. Domen, Chem. Mater. 20 (2008) 110.
acknowledge Prof. Alain Krief for his kind suggestion for the man-
[8] X. Zong, H.J. Yan, G.P. Wu, G.J. Ma, F.Y. Wen, L. Wang, C. Li, J. Am. Chem. Soc.
uscript revision. 130 (2008) 7176.
[9] T. Sakata, K. Hashimoto, T. Kawai, J. Phys. Chem. 88 (1984) 5214.
[10] K. Maeda, K. Teramura, N. Saito, Y. Inoue, K. Domen, J. Catal. 243 (2006) 303.
Appendix A. Supplementary material [11] J. Lee, H.H. Seliger, J. Chem. Phys. 40 (1964) 519.
[12] G. Ma, H. Yan, J. Shi, X. Zong, Z. Lei, C. Li, J. Catal. 260 (2008) 134.
Supplementary data associated with this article can be found, in [13] J.F. Moulder, W.F. Stickle, P.E. Sobol, K.D. Bomben, Handbook of X-ray
Photoelectron Spectroscopy, Perkin–Elmer Corp., Physical Electronics
the online version, at doi:10.1016/j.jcat.2009.06.024. Division, USA, 1992.
[14] J.A. Rodriguez, T. Jirsak, S. Chaturvedi, J. Chem. Phys. 110 (1999) 3138.
References [15] Y. Boudeville, F. Figueras, M. Forissier, J. Portefaix, J.C. Vedrine, J. Catal. 58
(1979) 52.
[16] T. Kida, G.Q. Guan, Y. Minami, T.L. Ma, A. Yoshida, J. Mater. Chem. 13 (2003)
[1] K. Maeda, K. Teramura, D.L. Lu, T. Takata, N. Saito, Y. Inoue, K. Domen, Nature
1186.
440 (2006) 295.
[17] I.J. Ferrer, P. Díaz-Chao, A. Pascual, C. Sánchez, Thin Solid Films 515 (2007)
[2] F.E. Osterloh, Chem. Mater. 20 (2008) 35.
5783.
[3] A. Kudo, Y. Miseki, Chem. Soc. Rev. 38 (2009) 253.
[18] K. Sayama, A. Nomura, T. Arai, T. Sugita, R. Abe, M. Yanagida, T. Oi, Y. Iwasaki, Y.
[4] H. Kato, K. Asakura, A. Kudo, J. Am. Chem. Soc. 125 (2003) 3082.
Abe, H. Sugihara, J. Phys. Chem. B 110 (2006) 9.
[5] I. Tsuji, H. Kato, A. Kudo, Angew. Chem. Int. Ed. 44 (2005) 3565.