CdS nanoparticles exhibiting quantum size effect by dispersion on tio 2: photocatalytic h2 evolution and photoelectrochemical measurements

Article abstract of DOI:10.1246/bcsj.82.528

Uniformly sized CdS nanoparticles exhibiting quantum size effect have been synthesized as a suspension in hexane solution containing oleic acid and 9-octadecenylamine as protective ligands. Our study focuses on utilizing the intrinsic optical property of

Full text of DOI:10.1246/bcsj.82.528

5
28
Bull. Chem. Soc. Jpn. Vol. 82, No. 4, 528–535 (2009)
© 2009 The Chemical Society of Japan
CdS Nanoparticles Exhibiting Quantum Size Effect by
Dispersion on TiO : Photocatalytic H Evolution
2
2
and Photoelectrochemical Measurements
1
1
2
3
3
Kiyonori Ogisu, Kazuhiro Takanabe, Daling Lu, Masaki Saruyama, Takahiro Ikeda,
3
3
1
Masayuki Kanehara, Toshiharu Teranishi, and Kazunari Domen*
1Department of Chemical System Engineering, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656
2Chemical Resources Laboratory, Tokyo Institute of Technology, 4259 Nagatsuta, Midori-ku, Yokohama 226-8503
3Graduate School of Pure and Applied Sciences, University of Tsukuba, Tennodai, Tsukuba 305-8571
Received November 20, 2008; E-mail: domen@chemsys.t.u-tokyo.ac.jp
Uniformly sized CdS nanoparticles exhibiting quantum size effect have been synthesized as a suspension in hexane
solution containing oleic acid and 9-octadecenylamine as protective ligands. Our study focuses on utilizing the intrinsic
optical property of the nanoparticles for photocatalytic H2 evolution and photoelectrochemical reaction by immobilizing
oxide substrate and removing protective ligands. Immobilized CdS nanoparticles on TiO2 were found to maintain their
original particle size and optical property after impregnation and removal of the protective ligands by oxidative heat
treatment and/or alkali treatment. The photocatalytic H2 evolution rates in Na2S­Na2SO3 solution under visible light were
strongly affected by the CdS particle size (quantum size effect), CdS loading amount, and TiO2 surface area. High
photoanodic current was observed for CdS/TiO2 sprayed on a fluoride-doped tin oxide (FTO) electrode, functioning as an
n-type semiconductor. The knowledge acquired in this study can be extended for further development in the field of
nanoparticle materials covered with protective ligands.
Hydrogen production from water using a semiconductor
photocatalyst is a key technology for achieving sustainable
particle with radius R is a function of the band gap energy Eg of
bulk semiconductor and kinetic energy (eq 1). The kinetic
energy can be calculated as the energy of a particle in a box
having a spherically symmetric square well potential of infinite
depth:
1
2
energy conversion. Some oxides, such as TiO , SrTiO , and
2
3
3
NaTaO3, function as highly efficient photoelectrodes and
photocatalysts under UV irradiation. To utilize sunlight how-
ever, a candidate material must possess appropriate visible light
absorption properties. Several chalcogenide materials, such as
ꢀ
ꢁ
2
hꢀ
2
E ¼ E þ
ln
g
ºln
ð1Þ
ꢀ
2
4
5
6
2m R
CdS, NaInS , and (AgIn) Zn
S , are promising photo-
2
2
x
2(1¹x)
catalysts because they exhibit appropriate absorption of visible
light.
where m* is the reduced effective mass of the conduction band
electron and valence band hole. º is the n-th root of the
ln
2
1
Among these chalcogenide materials, CdS has received
particular attention as a photocatalyst in recent years. It has
potential as an efficient photocatalyst for water splitting
because of its visible-light absorption (B.G. = 2.4 eV) and
the position of its band edges. The material itself however, is
unable to split H O to form H2 and O , and the material
spherical Bessel function of l-th order. The equation indicates
that the band gap of nanosize CdS (¯6 nm) is larger than that of
CdS bulk, and the gap increases with decreasing particle size.
Thus, nanoparticles show distinct oxidative and reductive
abilities for photocatalytic reactions. Decreasing particle size
also leads to an increase in surface area, altering mechanical,
thermal, and catalytic properties.
2
2
decomposes in standard aqueous solution during photoreac-
7­9
tion. To enable the use of this material, many researchers
have demonstrated the hydrogen evolution reaction in reducing
We synthesized well-tuned CdS nanoparticles with a very
narrow size-distribution in the presence of protective ligands
agents such as EDTA, S2 , SO , or phosphinate ions to
¹
2¹
suspended in organic solution.
22­24
Several studies have
3
prevent photoanodic dissolution.7 Stabilization of CdS on
­13
assessed the use of quantum dot CdS sensitized TiO2 as
photocatalyst material. In contrast to many other studies, where
CdS was synthesized from Cd solution and immobilized on a
support material simultaneously by in situ electrochemical
support materials such as TiO , AgI, ZnO, or HgS has also
2
1
4­16
been investigated,
and may improve the efficiency of
charge separation and photoreaction efficiency.
2
5­32
Semiconductor particles with diameters on the order of
nanometers exhibit unique size-dependent properties, such as
the quantum size effect which drastically alters electrical
methods or photodeposition,
we immobilized already-
prepared CdS nanoparticles on TiO to maintain the original
2
particle sizes and avoid aggregation upon removal of the
protective ligands and during photoreactions. The immobiliza-
tion of uniformly sized CdS nanoparticles in a controlled
1
7­20
In the conventional infinite-depth well model,21
properties.
the excitation energy level E of an ultra-small semiconductor
ln
Published on the web April 11, 2009; doi:10.1246/bcsj.82.528

K. Ogisu et al.
Bull. Chem. Soc. Jpn. Vol. 82, No. 4 (2009)
529
manner also enables us to elucidate details of particle size
effects, e.g., quantum size effects. By varying the immobiliza-
tion method, CdS particle size, surface area of the support, and
the amount of CdS and Pt (as a co-catalyst), the immobilized
CdS/TiO catalysts were tested for photocatalytic H evolution
Photocatalytic Reactions. H2 evolution reactions from 0.1 g
of CdS/TiO2 catalysts were performed using an aqueous solution
3
(
200 cm ) in the presence of 0.1 M Na2S and 0.1 M Na2SO3 as
sacrificial electron donors. A Pt co-catalyst was loaded by
photodeposition in the presence of the sacrificial electron donors
0.1 M Na2S and 0.1 M Na2SO3) using H2PtCl6¢6H2O (Aldrich;
2
2
(
²
and photoelectrochemical reactions in the presence of Na S/
2
37.50% as Pt).
Na2SO3 sacrificial reagent under visible light irradiation. This
study clearly shows the beneficial effects of immobilizing
nanoparticles with quantum size effects, providing high
efficiency and stability for photocatalytic reactions.
Photocatalytic reactions were carried out in a Pyrex reaction
vessel connected to a closed gas circulation system. The reaction
solution was evacuated several times to remove residual air from
the reaction solution, and then irradiated under visible light using a
3
00 W xenon lamp via a cutoff filter (­ ² 420 nm) to eliminate
Experimental
ultraviolet (UV) light and a water filter to block infrared light. The
amounts of H2 evolved were measured by gas chromatography
(GC-8A, Shimadzu; MS-5A column, TCD, Ar carrier). The
catalyst is denoted as Pt(n)/M wt % CdS(D)/TiO2, where n, M,
and D are the wt % of Pt based on CdS, wt % of CdS based on
TiO2, and particle size in nm. P25 was used as a TiO2 support
unless otherwise noted.
Immobilization of CdS Nanoparticles on TiO2 Support.
CdS nanoparticles were synthesized based on a previously reported
The CdS nanoparticles were immobilized on TiO2
photocatalysts then prepared by heat treatment and alkali
treatment. CdS nanoparticles (1.6 « 0.2, 2.9 « 0.3, 3.4 « 0.4,
.1 « 0.6, and 12.5 « 1.6 nm) suspended in hexane solution
Kanto Chemical, 96%) with oleic acid (bp 468 K) and oleyl
amine (bp 441 K) as protective ligands were added to TiO2
Degussa, P25) and mixed by ultrasonic agitation until the hexane
was vaporized at 298 K. In the heat treatment, dried samples were
heated at 473 or 623 K for 5 h in air. In the alkali treatment method,
method.2
2­24
4
(
Results and Discussion
(
Characterization of CdS Nanoparticles. The different
sized CdS nanoparticles (1.6­12.5 nm) protected by (Z)-9-
octadecenylamine and oleic acid were prepared according to
¹3
dried samples were added to 50 mL of 0.1 M (1 M = 1 mol dm
)
22­24
literature methods.
Figure 1 shows the TEM images of
NaOH solution and stirred for 1 h to remove protective ligands.
The samples were filtered and washed with 50 mL of distilled
water, then dried at 343 K in static air.
these CdS nanoparticles. All CdS nanoparticles were mono-
disperse and soluble in nonpolar solvents, such as hexane,
toluene, and chloroform. As shown in Figure 2, UV­vis spectra
Characterization. Differential thermal analysis and thermo-
gravimetry (TG-DTA; TG 8120, Rigaku) were used to study the
crystallization, decomposition, melting, and sublimation processes.
The crystalline structure of the catalyst was examined by powder
X-ray diffraction (XRD; Panalytical X’Pert Pro MPD). The
optical properties were analyzed with an ultraviolet­visible diffuse
reflectance spectrometer (DRS; V-560, Jasco). Particle size and
photocatalyst morphology were examined by transmission electron
microscopy (TEM; JEM-2010F, JEOL). The elemental composi-
tions were probed by X-ray fluorescence (XRF; Element Analyzer
JSX-3202C, JEOL) and energy-dispersive X-ray spectroscopy
a
b
(
EDX; Emax-7000, Horiba). The Brunauer­Emmett­Teller (BET)
specific surface area was measured using a Coulter SA-3100
instrument at 77 K.
c
d
CdS/TiO2 electrodes were fabricated by spreading a viscous
slurry of the prepared powder on fluorine-doped tin oxide (FTO)
transparent conductive glass (ca. 20 ³; Asahi Glass). Before
spreading the slurry, the substrates were ultrasonicated, first in
acetone and then in distilled water. The slurry consisted of the
sample (0.1 g), water (100 mL), acetylacetone (100 mL), and 10 mL
of surfactant (Triton X-100, Aldrich). The electrode (coated area:
2
1
© 4 cm ) was then heated in air at 573 K for 1 h to remove
e
organic solvent and surfactant and to improve the contact between
catalyst and substrate.
Current­voltage and current­time curves were measured using a
conventional Pyrex electrochemical cell attached to a planar side
window. The cell consisted of a prepared electrode, a platinum wire
counter electrode (diameter: 1 mm, length: 250 mm), and an
Ag/AgCl reference electrode. Aqueous solutions of 0.1 M Na2S­
Na2SO3 were used as electrolytes. The electrolytes were saturated
with argon prior to each electrochemical measurement, and elec-
trode potential was controlled by a potentiostat (HZ-5000, Hokuto
Denko or SDPS-501C, Syrinx). Light irradiation was performed
using a 300 W xenon lamp with a cutoff filter (­ ² 420 nm).
Figure 1. TEM images of CdS nanoparticles with (a)
1.6 « 0.2, (b) 2.9 « 0.3, (c) 3.4 « 0.4, (d) 4.1 « 0.6, and
(e) 12.5 « 1.6 nm in size. The scale bars indicate 50 nm.

5
30
Bull. Chem. Soc. Jpn. Vol. 82, No. 4 (2009)
Photocatalytic Performance of CdS-QD on TiO2
5
0
0
c
b
a
0
2
DTA
-
-
TG
-
50
4
0
5
0
DTA
0
-
-
2
TG
-
50
4
0
300
350
400
450
500
550
600
5
0
Wavelength / nm
DTA
Figure 2. UV­vis spectra of n-hexane solution of CdS
nanoparticles with (a) 1.6 « 0.2, (b) 2.9 « 0.3, (c) 3.4 «
0
-
-
2
TG
0.4, (d) 4.1 « 0.6, and (e) 12.5 « 1.6 nm in size.
-50
4
300
of 1.6, 2.9, 3.4, and 4.1 nm sized CdS nanoparticles showed
exiton peaks at 370, 432, 449, and 457 nm, respectively, due to
the quantum size effect, whereas 12.5 nm CdS nanoparticles
showed bulk like spectrum.
500
Temperature/ K
700
900
Figure 3. TG-DTA curves under dry flowing air for 0.072
wt % CdS(3.4 « 0.4)/TiO2 catalysts with (a) no treatment,
b) heat treatment (HT) at 473 K for 5 h, and (c) alkali
treatment (AT) in 0.1 M NaOH solution at 298 K for 1 h.
Immobilization of CdS Nanoparticles on TiO Support.
2
(
Immobilization of CdS nanoparticles on TiO2 support was
accomplished by sonicating TiO suspension in CdS hexane
2
solution at room temperature. Hexane proved to be the most
effective solvent in supporting as many CdS nanoparticles as
possible on TiO owing to the lowest polarity.
2
Removal of the Protective Ligand. The prepared CdS
nanoparticles were stabilized in the presence of protective
ligands, oleic acid (bp 468 K), and oleyl amine (bp 441 K). We
first attempted to remove these protective ligands to improve
the accessibility of the CdS surfaces. Figure 3 shows the TG-
DTA curves under dry flowing air for 0.072 wt % CdS(3.4)/
f
b
TiO catalysts with (a) no treatment, (b) heat treatment (HT) at
2
4
73 K for 5 h, and (c) alkali treatment (AT) in 0.1 M NaOH
d
c
solution at 298 K for 1 h. The TG profile for non-treated
CdS/TiO2 (Figure 3a) shows weight loss at 500­700 K,
presumably ascribed to the removal of protective ligands such
as 9-octadecenylamine and oleic acid and the decomposition
of sulfide to oxide. This is consistent with the observed
a
e
3
00
400
500
600
700
Wavelength / nm
exothermic profile in the corresponding temperature range, as
shown in the DTA curve.3
3­35
Bulk CdS showed no weight
Figure 4. UV­vis DR spectra for (a) TiO (P25), (b) bulk
2
change at 300­900 K (results not shown). Figures 3b and 3c
show that HT at 473 K for 5 h and AT drastically reduced the
degree of weight loss and exothermicity.
CdS (99.999%, Mitsuwa Chemical), CdS(3.4 « 0.4)/TiO2
samples with (c) no treatment, (d) HT at 473 K for 5 h,
(e) HT at 623 K for 5 h, and (f) AT in 0.1 M NaOH.
Figure 4 shows UV­vis DRS for (a) TiO (P25) and (b) CdS
2
(
Mitsuwa) for reference, and for 0.072 wt % CdS(3.4)/TiO2
samples with (c) no treatment, (d) HT at 473 K for 5 h, (e) HT at
23 K for 5 h, and (f) AT in 0.1 M NaOH. The absorption band
edges of TiO and CdS (Figures 4a and 4b) were 380 and
nanoparticles while maintaining the particle size. The absorp-
tion at wavelengths beyond the absorption edge of CdS, which
6
was observed for CdS(3.4)/TiO by HT at 473 K, was probably
2
due to reduced Cd species. On the contrary, the absorption by
CdS nanoparticles completely vanished, and only adsorption
due to TiO2 was observed after HT at 623 K for 5 h (e).
Therefore, HT at 473 K and AT methods were used to remove
the protective ligands.
2
5
30 nm corresponding to B.G. of 3.26 and 2.34 eV, respective-
ly. The absorption peak of the as-prepared CdS nanoparticle
3.4 « 0.4 nm) was 449 nm (B.G. = 2.76 eV), a shorter wave-
length by about 90 nm than for bulk CdS (blue shift,
B.G. = 0.46 eV) due to the quantum size effect. The observed
peaks for 0.072 wt % CdS(3.4)/TiO with (c) no treatment,
(
¦
Loading of CdS Nanoparticle. Figure 5 shows UV­vis
DRS for 0.072 wt % CdS/TiO prepared with different CdS
2
2
(
d) HT method at 473 K, and (f) AT method remained almost
particle sizes (1.6 « 0.2 (c), 2.9 « 0.3 (d), 3.4 « 0.4 (e),
4.1 « 0.6 (f), and 12.5 « 1.6 nm (g)) by HT at 473 K for 5 h
unchanged from that of as-prepared CdS nanoparticles,
confirming that the methods successfully immobilized CdS
as well as (a) TiO (P25) and (b) bulk CdS (Mitsuwa). No
2

K. Ogisu et al.
Bull. Chem. Soc. Jpn. Vol. 82, No. 4 (2009)
531
d
f
b
e
f
b
g
c
g
b
a
e
4
00
450
500
550
Wavelength / nm
d
d
e
f
c
g
a
c
a
3
00
400
500
600
3
00
400
Wavelength / nm
500
Wavelength / nm
Figure 6. UV­vis DR spectra for (a) TiO2 (P25), (b) bulk
CdS (99.999%, Mitsuwa Chemical), CdS nanoparticles
loaded on TiO2 after HT treatment; The amounts of CdS
were (c) 0.072, (d) 0.3, (e) 0.7, (f) 1, and (g) 3 wt %.
Figure 5. UV­vis DR spectra of (a) TiO2 (P25), (b) bulk
CdS (99.999%, Mitsuwa Chemical), several particle sizes
(
1.6 « 0.2 (c), 2.9 « 0.3 (d), 3.4 « 0.4 (e), 4.1 « 0.6 (f),
and 12.5 « 1.6 nm (g)) CdS nanoparticles on TiO2 by HT
method.
Table 1. The Rate of H2 Evolution from TiO2 (Support),
CdS Nanoparticle, CdS/TiO2 (No Treatment), CdS/TiO2
absorption due to CdS was observed for CdS(1.6)/TiO2
(HT or AT), and Pt/CdS/TiO (HT or AT) in a 200 mL of
2
(Figure 5c), which originally showed absorption at 370 nm
Solution Containing 0.1 M Na S­Na SO , 0.1 g of Catalyst
2
2
3
before immobilization. In the case of CdS(1.6)/TiO , the
(CdS: 0.072 wt %, Pt: 50 wt % for CdS) under Irradiation at
­ > 420 nm
2
absorption above 370 nm was subtly observed whereas the
absorption for the unsupported CdS(1.6) did not have absorp-
tion in this range. This result indicates that a portion of
the CdS(1.6) particles slightly sintered; yet the major por-
tion of CdS(1.6) is still believed to remain constant in
size. The CdS(2.9)/TiO , CdS(3.4)/TiO , and CdS(4.1)/TiO
¹1
Sample
Rate of H2 evolution/µmol h
TiO2
0
1
23
68
104
119
CdS nanoparticle
CdS/TiO2(HT)
CdS/TiO2(AT)
Pt/CdS/TiO2(HT)
Pt/CdS/TiO2(AT)
2
2
2
(Figures 5d­5f) showed absorption at 430, 446, and 459 nm,
respectively. These absorptions corresponded well to those of
the original CdS nanoparticles at 432, 449, and 457 nm for CdS
particle sizes of 2.9, 3.4, and 4.1 nm, respectively.2
2­24
As seen
in Figure 5g, the CdS(12.5)/TiO absorption edge was ob-
support only (B.G. = 3.26 eV), 72 µg of CdS nanoparticle
colloid without support (B.G. = 2.76 eV), 0.1 g of 0.072 wt %
CdS(3.4)/TiO2 (HT or AT), and 0.1 g of Pt(50) loaded
0.072 wt % CdS(3.4)/TiO . No H evolution was detected
2
served at 540 nm, matching that of the original CdS nano-
particle (12.5 « 1.6 nm).2
2­24
Therefore, it can be concluded
that the optical properties and particle size of CdS supported on
2
2
TiO remained unchanged from the original CdS particles.
from TiO support only, due to its obvious lack of visible
2
2
We attempted to increase the CdS loading on TiO2 by
starting with CdS with a particle size of 3.4 « 0.5 nm. Figure 6
light response. Although protective ligands should have been
removed from CdS particles in Na S/Na SO solution (pH 13),
2
2
3
shows UV­vis DRS for CdS/TiO (0.072­3 wt % CdS) after
only negligible H was evolved from CdS nanoparticle colloid
2
2
HT. The CdS/TiO samples after AT had very similar UV­vis
(without support), probably because unstable CdS nanoparti-
cles may aggregate easily when they are not immobilized on a
support. Supported CdS nanoparticles, on the other hand, had
much higher photocatalytic activity (>20 times) than CdS
nanoparticles only, although the amounts of CdS in the reaction
solution were identical. Moreover, the deposition of Pt on CdS/
2
DRS properties (Figure S1, Supporting Information). Absorp-
tion was observed at 450 nm in the range 0.072­1 wt % CdS
loading (Figure 6). The absorption edge shifted to longer
wavelengths with increasing amount of CdS (1­3 wt %). This
indicates that the CdS particles aggregated to form larger
particles, where the quantum size effect was no longer
operative. Assuming the CdS particles were spherical or square
in shape, 14.7 or 11.6 wt % of CdS(3.4) could be theoretically
TiO considerably enhanced the photocatalytic activity (µ5
2
fold). The photocatalytic activity for CdS/TiO (AT) was about
2
three times higher than that for CdS/TiO2 (HT). Figure 4
shows that the absorption at the wavelengths beyond the
absorption edge (­ > 460 nm) was observed in the case of
2
¹1
loaded to form a CdS monolayer on TiO (P25, 54 m g ),
2
respectively. A surface coverage of 6.8% (sphere) or 8.6%
(
square) was estimated as the loading at which CdS nano-
particles should begin to aggregate in the current treatments.
Photocatalytic H2 Evolution Reaction. Table 1 lists
the steady-state H evolution rates from 0.1 g of TiO (P25)
CdS/TiO (HT), most likely due to the existence of sulfur
defects on the surface. The lack of absorption in this range for
2
CdS/TiO (AT) may indicate the defect sites are much less than
2
those for CdS/TiO (HT), probably accounting for the dif-
2
2
2

5
32
Bull. Chem. Soc. Jpn. Vol. 82, No. 4 (2009)
Photocatalytic Performance of CdS-QD on TiO2
Table 2. The Rate of H2 Evolution for Several Particle Sizes
of CdS (1.6 « 0.2, 2.9 « 0.3, 3.4 « 0.4, 4.1 « 0.6, and
450
4
3
00
50
12.5 « 1.6 nm) Loaded on TiO2
a)
b)
Diameter Band gap Surface area
H2 evolution rate
2
¹1
¹1
300
/nm
/eV
/m g
/µmol h
2
2
1
50
00
50
1
2
3
4
.6 « 0.2
.9 « 0.3
.4 « 0.4
.1 « 0.6
3.35
2.87
2.76
2.71
2.34
778
429
356
304
89
43(346)
60(®)
104(227)
66(106)
2(6)
1
2.5 « 1.6
100
50
0
a) Surface areas are estimated by particle size in assumption to
sphere in shape. b) This value in parenthesis is the rate of H2
evolution under light irradiation (­ > 300 nm). In the case of
CdS(2.9)/TiO2 was not measured.
0
20 40 60 80 100
2 -1
Support BET surface area / m g
Figure 7. Dependence of the rate of H2 evolution from Pt
100 wt %)/CdS (( ) 0.072 and ( ) 0.3 wt %, particle
size = 3.4 nm)/TiO2 by HT upon the BET surface area;
ferences in photocatalytic performance. In the case of Pt-
deposited samples however, the photocatalytic activity was
similar in CdS/TiO (HT) and CdS/TiO (AT), which may
indicate that the Pt gives preferential surface sites that do not
depend on the CdS defect sites. These results clearly show that
immobilization of nanoparticles on a support material is
essential to establish their intrinsic activity for photocatalytic
reaction.
(
2
2
2
¹1
support candidates were rutile (SBET = 4 m g ), anatase
2
¹1
2 ¹1
(
(
SBET = 29 m g ), anatase (SBET = 103 m g ), and P25
SBET = 54 m g¹1).
2
Next, the effects of the TiO support surface area on the
2
Unusual properties due to quantum size effects should
significantly influence photocatalytic activity, and can be
probed by varying CdS particle size. Table 2 shows the H2
evolution rates of the Pt(50)/0.072 wt % CdS/TiO2 with
starting CdS particle sizes of 1.6 « 0.2, 2.9 « 0.3, 3.4 « 0.4,
photocatalytic activity were investigated. TiO support candi-
2
2
¹1
2 ¹1
dates were rutile (SBET = 4 m g ), anatase (S
= 29 m g ),
BET
2
¹1
anatase (SBET = 103 m g ), and P25 (ca. 70% anatase, 30%
2
¹1
rutile, SBET = 54 m g ). The H evolution rates for the
2
CdS(3.4)/TiO catalysts are plotted against BET surface area
2
4
.1 « 0.6, and 12.5 « 1.6 nm under visible light and UV light
in Figure 7. Without support (plots at x = 0 in Figure 7), the
photocatalytic activity for CdS nanoparticles was almost
negligible, clearly demonstrating the positive effects of CdS
nanoparticle dispersion onto the support. Even for the low
irradiation. Under visible light irradiation, the photocatalytic
activity increased as particle size increased up to 3.4 nm,
but then began to decrease. Under UV-light irradiation, the
photocatalytic activity increased with decreasing particle size.
The band gap of CdS nanoparticles measured by UV­vis DRS
loading of 0.072 wt %, the H evolution rate was lower for a
2
low-surface-area rutile sample than for other samples with
higher surface areas. The comparable rates for the latter
samples indicate that nanoparticles were well-dispersed on the
support at low CdS loading (0.072 wt %). For 0.3 wt % CdS
loaded samples, the photocatalytic activity increased with
(Figure 5) and CdS surface area estimated from the particle size
assuming a spherical shape are summarized in Table 2. With
increasing particle size, the band gaps narrowed due to
quantum size effects. On the other hand, the estimated CdS
surface area drastically decreases with increasing particle size.
It should be noted that 12.5 nm CdS particles exhibited the
band gap of bulk CdS material. The photocatalytic activity
should be affected by absorptions of visible and UV light
by the catalysts, and the ease of oxidation and reduction
determined by the positions of the conduction and valence
bands of the catalysts. Under UV light irradiation, the H2
evolution rate decreased with increasing initial CdS particle
size, corresponding to a decrease in band gap (i.e., oxidation­
reduction ability) and an increase in surface area. Under visible
light irradiation, CdS samples with 1.6 nm particles did not
have sufficient absorption in the visible light region (absorp-
2
¹1
increasing the surface area of the support by 54 m g . It is
clear that a higher surface area support can disperse the active
CdS species more efficiently, resulting in enhancement of the
H evolution rate. 0.3 wt % CdS/TiO samples showed higher
2
2
activity than 0.072 wt % CdS/TiO samples. Dependence of the
2
rate of H evolution upon the amount of Pt cocatalyst loaded on
2
0.072 wt % CdS(3.4)/TiO2 are shown in Figure S2. The
Pt(100)-loaded photocatalyst showed high activity.
The H2 evolution rate is plotted against the amount of
CdS(3.4) in HT (a) or AT (b) in Figure 8. The H evolution
2
rates for the catalysts obtained by HT or AT reached maxima
when 0.7 or 0.3 wt % of CdS was loaded, respectively. The rate
drastically decreased for highly loaded (1 wt %) samples. UV­
vis spectra (Figure 6) showed that the absorption shifted to
wider wavelength (narrower band gap) compared to samples
with lower loading, indicating that particle size increased with
increasing amount of CdS, while photocatalytic rates decreased
accordingly. These results suggest that the immobilization of
CdS nanoparticles exhibiting quantum size effects resulted in
tion: 370 nm), resulting in a comparatively low H evolution
2
rate. By increasing CdS particle size beyond 3.4 nm, the
photocatalytic activity decreased because of a depression of the
oxidative­reductive power with a narrower band gap. It is
notable that nanoparticles exhibiting quantum size effects gave
much higher rates than the 12.5 nm CdS sample, which had
bulk CdS nature, suggesting that utilization of these nano-
particles had significant benefits for this reaction.
high-efficiency H evolution.
2

K. Ogisu et al.
Bull. Chem. Soc. Jpn. Vol. 82, No. 4 (2009)
533
6
5
4
3
2
1
00
1
1
200
000
00
00
00
00
00
0
a
b
8
00
600
400
2
00
0
3
00
400
500
600
700
0
.0 0.2 0.4 0.6 0.8 1.0
Loading amount of CdS (wt%)
Wavelength / nm
Figure 9. Wavelength dependence of the steady rate of H2
evolution and UV­vis DRS results for Pt (100 wt % for
CdS)/CdS (0.3 wt %)/TiO2 (P25) with AT followed by HT
in a 200 mL of aqueous solution containing 0.1 M Na2S­
Na2SO3, 0.1 g of catalyst under visible irradiation at
Figure 8. Dependence of the rate of H2 evolution on the
loading amount of CdS in HT (a) or AT (b); Catalyst
(
(
0.1 g), Pt (100 wt % for CdS)/CdS (0.072­1 wt %)/TiO2
P25); 0.1 M Na2S­Na2SO3 aqueous solution under visible
irradiation at ­ ² 420 nm.
­
² 420 nm.
Finally, we carried out photocatalytic measurements of
Pt(100)/0.3 wt % CdS(3.4)/TiO2 with AT followed by HT at
8
7
6
5
4
3
2
1
0
4
73 K under visible light irradiation in the presence of Na S­
2
a
Na SO . This procedure takes advantage of both treatments.
2
3
The alkali treatment ensures the removal of part of the
protective ligands while maintaining the CdS particle size. The
remaining protective ligands after the alkali treatment can be
further removed by the successive heat treatment with lower
exotherm by combustion, thus suppressing creation of sulfur
defects than by the direct heat treatment procedure without the
alkali treatment.
200
b
1
1
50
00
50
¹
1
The rate of H2 evolution was 616 µmol h , the highest
value observed in this study, and remained constant for
0
-
1.2
-0.8
-0.4
0.0
Potential / V (vs. Ag/AgCl)
5
h (Figure S3 in Supporting Information). Figure 9 shows
the wavelength dependence of photocatalytic H2 evolution
from aqueous Na S­Na SO solution over Pt(100)/0.3 wt %
b
2
2
3
-
1.5 -1.2 -0.9 -0.6 -0.3 0.0
Potential / V (vs. Ag / AgCl)
CdS(3.4)/TiO catalyst. The H evolution rate decreased with
2
2
increasing cutoff wavelength, corresponding to excitation at the
absorption edge of CdS nanoparticles (3.4 « 0.4 nm). We
confirmed that these photocatalytic H2 evolution reactions
occur by the photogeneration of electrons and holes under
visible light irradiation in the presence of Na S­Na SO .
Figure 10. Photocurrent versus voltage of CdS (0.3 wt %)/
TiO2 (P25) after AT followed by HT on FTO in 0.1 M
Na2S­Na2SO3 solution at ¹0.2 V versus Ag/AgCl under
visible irradiation at ­ ² 420 nm.
2
2
3
Photoelectrochemical Properties. Figure 10 shows the
current­voltage curves for 0.3 wt % CdS(3.4)/TiO2 sprayed
on an FTO electrode compared to bulk CdS (99.999%,
Mitsuwa Chemical) in a solution containing Na2S and Na2SO3
as sacrificial electron donors (pH 13) under visible light
the difference in band gap mainly originated from the differ-
ence in conduction edges, and the positions of the valence
bands of CdS(3.4)/TiO and bulk CdS/TiO were very similar,
2
2
1
7
(
­ > 420 nm). The measurements were conducted by scanning
as reported previously. The large difference in photocatalytic
activity for H evolution between quantum dot CdS(3.4)/TiO
potential from ¹1.5 to 0 V versus Ag/AgCl. The observed
2
2
anodic photocurrent is attributed to oxidation of Na S and
and bulk-like CdS(12.5)/TiO as discussed above (Table 2),
2
2
Na SO , indicating that the CdS(3.4)/TiO functioned as an
n-type semiconductor (Figure S4 in Supporting Information).
A photocurrent onset for CdS(3.4)/TiO2 was observed at
can be attributed to the high reduction capability of quantum
dot CdS derived from a negatively shifted conduction edge. In
2
3
2
addition, a large photocurrent was observed for CdS/TiO
2
¹2
¹1.45 V; a ¹0.37 V negative shift compared to bulk CdS
(6.8 mA cm at ¹0.2 V versus Ag/AgCl) compared to bulk
¹2
electrode (¹0.108 V). From UV­vis DRS, the band gap of
CdS(3.4)/TiO and bulk CdS/TiO were determined to be 2.76
CdS (0.1 mA cm at ¹0.2 V versus Ag/AgCl) under visible
¹2
light irradiation with the power density 0.49 W cm (­ >
420 nm), reflects the high reduction capability of quantum dot
CdS. The effect may be attributed to charge separation by TiO2
2
2
and 2.34 eV, respectively, and the difference in band gap was
.42 eV, very similar to the difference of onset potential. Thus
0

5
34
Bull. Chem. Soc. Jpn. Vol. 82, No. 4 (2009)
Photocatalytic Performance of CdS-QD on TiO2
+
reduction of H to H under visible light irradiation in the
2
1
1
1
4
2
0
8
6
4
2
0
presence of a sacrificial electron donor (Na S­Na SO ).
2
2
3
Quantum size effects in CdS particles on TiO were observed
2
from 1.6 to 4.1 nm (1.6 « 0.2, 2.9 « 0.3, 3.4 « 0.4, and
4
.1 « 0.6 nm), and the photocatalytic activity was maximum
at 3.4 « 0.4 nm, about fifty times higher than 12.5 « 1.6 nm
CdS with similar properties to bulk CdS. To further investigate
the preparation method, the loading amount of CdS or Pt and
several supports with different surface areas were examined.
¹1
The photocatalytic activity was 616 µmol h at Pt (100 wt %
2
¹1
for CdS)/CdS (0.3 wt %)/TiO (P25, S
= 54 m g ) after
BET
2
alkali treatment followed by heat treatment. For CdS/TiO
on an FTO electrode, photoanodic current (6.8 mA cm )
2
¹2
4
00
440
480
520
560
was measured in the presence of 0.1 M Na S­Na SO under
2
2
3
Wavelength / nm
¹2
visible light irradiation with the power density 0.49 W cm
­ > 420 nm). It was shown that CdS functions as a high
(
Figure 11. Wavelength dependence of IPCE and UV­vis
DRS results for CdS (0.3 wt %)/TiO2 (P25) with AT
followed by HT on FTO in 0.1 M Na2S­Na2SO3 solution
at ¹0.2 V versus Ag/AgCl under visible irradiation at
activity n-type semiconductor. The high efficiency was
attributed not only to quantum size effects, but also to charge
separation due to the difference between the bottoms of the
conduction bands of CdS and TiO . There is still a large margin
­
² 420 nm.
2
for photocatalytic activity improvement by optimizing type
and surface area of the support with supplementary deposition
of CdS. The knowledge acquired in this study can be applied
for general usage of organic-ligand-protected nanoparticle
materials.
and the migration without barriers between quantum dot CdS
and TiO2 support when photoexcited electrons transfer to the
2
9,31
TiO support.
There is a possibility that the CdS particles
2
on the grain boundary between TiO particles may cause loss of
2
the photoanodic currents. We cannot rule out that the currents
may be improved by a further optimized preparation method of
This work was supported in part by Global COE Program
(Chemistry Innovation through Cooperation of Science and
Engineering), MEXT, Japan.
the electrode: for instance, the synthesis of the TiO thin layer
2
prior to CdS introduction should avoid the resistance due to
CdS present between the grain boundaries of TiO , thus
improving photoelectrochemical performance.
2
Supporting Information
Figure 11 shows the dependence of IPCE (¹0.2 V versus
Ag/AgCl) on the cutoff wavelength of incident light for the
UV­vis DR spectra for CdS/TiO2 after AT (Figure S-1). Time
course of H2 evolution of several aliquots of Pt loaded on CdS/
TiO2 (Figure S-2). Time course of H2 evolution for Pt(100)/
0
.3 wt % CdS/TiO electrode. The incident photon-to-photo-
2
0
.3 wt % CdS(3.4)/TiO2 with AT followed by HT (Figure S-3).
current efficiency, IPCE, is defined by the following eq 2.
(a) Current­time at ¹0.2 V vs. Ag/AgCl and (b) current­potential
hc iphoto
curves for 0.3 wt % CdS(3.4)/TiO2 on FTO electrode (Figure S-4).
This material is available free of charge on the Web at: http://
www.csj.jp/journals/bcsj/.
IPCEð%Þ ¼ e ­ꢁ ꢁ 100
ð2Þ
¹
2
where iphoto, ­, and ¯ are photocurrent density (A m ),
¹2
wavelength (m), and light flux (W m ), and c, h, and e
8
¹1
References
are the speed of light (2.998 © 10 m s ), Plank’s constant
¹
34
1
2
3
4
5
A. Fujishima, K. Honda, Nature 1972, 238, 37.
(
(
6.626 © 10 W s), and the elementary electric charge
¹
19
K. Domen, A. Kudo, T. Onishi, J. Catal. 1986, 102, 92.
A. Kudo, H. Kato, Chem. Phys. Lett. 2000, 331, 373.
J. F. Reber, K. Meier, J. Phys. Chem. 1984, 88, 5903.
A. Kudo, A. Nagane, I. Tsuji, H. Kato, Chem. Lett. 2002,
1.602 © 10 A s). A monochromatic light source (Asahi
Spectra MAX-301) was used with a bandpath filter (half
bandwidth: «10 nm) in the wavelength region of 420 to
5
60 nm. The DRS results for the 0.3 wt % CdS(3.4)/TiO2
powder are also shown. Photocurrent was observed at up to
00 nm, and the variation in photocurrent with respect to
8
82.
6
I. Tsuji, H. Kato, H. Kobayashi, A. Kudo, J. Am. Chem.
5
Soc. 2004, 126, 13406.
wavelength was in good agreement with the DRS results. The
observed photocurrent can thus be attributed to a photoelec-
trochemical oxidation reaction via the band gap transition of
7
8
R. J. Williams, J. Chem. Phys. 1960, 32, 1505.
H. Gerischer, J. Electroanal. Chem. 1975, 58, 263.
9
J. R. Wilson, S.-M. Park, J. Electrochem. Soc. 1982, 129,
CdS. The IPCE at 420 nm for CdS/TiO at ¹0.2 V versus
2
1
1
49.
Ag/AgCl was thus derived to be 12.9%.
1
0
J. R. Darwent, J. Chem. Soc., Faraday Trans. 2 1981, 77,
J. R. Harbour, R. Wolkow, M. L. Hair, J. Phys. Chem. 1981,
703.
Conclusion
11
CdS nanoparticles dispersed on TiO to maintain particle
2
85, 4026.
size were successfully prepared by heat treatment and alkali
treatment to effectively remove protective ligands. The photo-
12 T. Inoue, T. Watanabe, A. Fujishima, K. Honda, K.
Kohayakawa, J. Electrochem. Soc. 1977, 124, 719.
13 H. Minoura, M. Tsuiki, Electrochim. Acta 1978, 23, 1377.
catalytic measurements suggest that CdS/TiO catalyzed the
2

K. Ogisu et al.
Bull. Chem. Soc. Jpn. Vol. 82, No. 4 (2009)
535
1
4
K. R. Gopidas, M. Bohorquez, P. V. Kamat, J. Phys. Chem.
106, 8967.
27 J. C. Yu, L. Wu, J. Lin, P. Li, Q. Li, Chem. Commun. 2003,
1552.
28 J. C. Tristão, F. Magalhães, P. Corio, M. T. C. Sansiviero,
J. Photochem. Photobiol., A 2006, 181, 152.
29 S. Chen, M. Paulose, C. Ruan, G. K. Mor, O. K. Varghese,
D. Kouzoudis, C. A. Grimes, J. Photochem. Photobiol., A 2006,
177, 177.
30 J. S. Jang, W. Li, S. H. Oh, J. S. Lee, Chem. Phys. Lett.
2006, 425, 278.
31 H. Park, W. Choi, M. R. Hoffmann, J. Mater. Chem. 2008,
18, 2379.
1
6
1
990, 94, 6435.
1
5
S. Hotchandani, P. V. Kamat, J. Phys. Chem. 1992, 96,
A. Eychmüller, A. Hässelbarth, H. J. Weller, J. Lumin.
834.
1
6
992, 53, 113.
17
18
19
20
21
L. E. Brus, J. Chem. Phys. 1984, 80, 4403.
Y. Nosaka, J. Phys. Chem. 1991, 95, 5054.
Y. Wang, Acc. Chem. Res. 1991, 24, 133.
A. Henglein, Chem. Rev. 1989, 89, 1861.
Al. L. Efros, A. L. Efros, Sov. Phys. Semicond. 1982, 16,
7
72.
2
2
T. Teranishi, M. Saruyama, M. Kanehara, Chem. Lett. 2009,
32 C.-F. Chi, Y.-L. Lee, H.-S. Weng, Nanotechnology 2008,
19, 125704.
33 P. J. Sebastian, M. E. Calixto, Thin Solid Films 2000, 360,
128.
34 F. El-Tantawy, K. M. Abdel-Kader, F. Kaneko, Y. K. Sung,
Eur. Polym. J. 2004, 40, 415.
35 L. Yang, Q. Shen, J. Zhou, K. Jiang, Mater. Lett. 2005, 59,
2889.
3
8, 194.
J. Joo, H. B. Na, T. Yu, J. H. Yu, Y. W. Kim, F. Wu, J. Z.
Zhang, T. Hyeon, J. Am. Chem. Soc. 2003, 125, 11100.
M. Kanehara, H. Arakawa, R. Hironaga, T. Teranishi,
Trans. Mater. Res. Soc. Jpn. 2006, 31, 437.
S. G. Hickey, D. J. Riley, E. J. Tull, J. Phys. Chem. B 2000,
04, 7623.
T. Hirai, Y. Bando, I. Komasawa, J. Phys. Chem. B 2002,
2
3
2
4
2
5
1
2
6