CuO/Pd composite photocathodes for photoelectrochemical hydrogen evolution reaction

ww.sciencedirect.com

i n t e rn a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 3 9 ( 2 0 1 4 ) 7 6 8 6e7 6 9 6

Available online at w

ScienceDirect

journal homepage: www.elsevier .com/locate/he

CuO/Pd composite photocathodes forphotoelectrochemical hydrogen evolution reaction

Xin Guo a, Peng Diao a,*, Di Xu a, Shan Huang a, Yang Yang a, Tao Jin a,Qingyong Wua, Min Xiang a, Mei Zhang b,*aKey Laboratory of Aerospace Materials and Performance (Ministry of Education), School of Materials Science and

Engineering, Beihang University, Beijing 100191, PR Chinab State Key Laboratory of Advanced Metallurgy, School of Metallurgical and Ecological Engineering, University of

Science and Technology Beijing, Beijing 100083, PR China

a r t i c l e i n f o

Article history:

Received 9 October 2013

Received in revised form

27 February 2014

Accepted 7 March 2014

Available online 13 April 2014

Keywords:

Photoelectrochemical cell

Solar water splitting

Hydrogen evolution reaction

Copper oxide

Palladium nanoparticles

* Corresponding authors. Tel./fax: þ86 01 823E-mail addresses: [email protected], pd

http://dx.doi.org/10.1016/j.ijhydene.2014.03.00360-3199/Copyright ª 2014, Hydrogen Energ

a b s t r a c t

CuO has been considered as a promising photocathodic material for photoelectrochemical

(PEC) hydrogen evolution reaction (HER). In this work, CuO films were prepared by a facile

and cost-effective method that involves solution synthesis, spin-coating and thermal

treatment processes. The resulting CuO films had a monoclinic crystal structure with

bandgap energy of 1.56 eV and a conduction band position of 3.73 eV below the vacuum

level in borate buffer solution. The CuO films exhibited good PEC activity toward HER and

the preparation conditions had great effect on the activity. The photoactivity of the CuO

film decayed to approximately 19% of its original value after reaction for 10 h under illu-

mination. The reduction of CuO to Cu2O has been confirmed to be a parallel competitive

reaction against HER. The mismatched band structure of the resulting CuO/Cu2O hetero-

junction was believed to be the main cause of the decay of photoactivity. The photo-

assisted electrodeposition method was developed to prepare CuO/Pd composite photo-

cathode. The presence of Pd on CuO greatly increased the photocurrent especially at low

overpotentials. In addition, the CuO/Pd composite exhibited significantly improved pho-

tostability compared to CuO. This work demonstrates the feasibility of increasing PEC

activity and stability of CuO-based photocathodes by combining CuO with noble metal

nanoparticles.

Copyright ª 2014, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights

reserved.

Introduction

The conversion of solar energy into chemical fuels is an

attractive, clean and sustainable solution to the growing en-

ergy problem [1]. Solar-driven water splitting based on semi-

conductor materials is an artificial photosynthesis process

[email protected] (P. Diao), zh84y Publications, LLC. Publ

that stores the solar energy in the chemical bonds of the two

products H2 and O2. Photoelectrochemical (PEC) cells are

widely used as solar-powered water splitting devices [2e4]

because they combine solar energy collection with water

electrolysis. Moreover, in PEC cells, the cathodic and the

anodic reactions occur at cathode and anode, respectively,

allowing the investigation of only one half-cell reaction. PEC

[email protected] (M. Zhang).

ished by Elsevier Ltd. All rights reserved.

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 3 9 ( 2 0 1 4 ) 7 6 8 6e7 6 9 6 7687

hydrogen evolution reaction (HER) is the cathodic half-cell

reaction which has attracted great research interests [5e9]

because H2 is an environmentally benign fuel that produces

only water upon combustion. PEC HER is usually based on a p-

type semiconductor/electrolyte interface, where the photo-

generated electrons are injected from semiconductor into

water by the field of space charge region to generate H2. In

principle, semiconductors with conduction band position

more negative than the potential of H2O/H2 can be used as

photocathode materials. To develop efficient photocathodes,

materials with small bandgap energy are preferred because

only the photons with higher energy than the bandgap of a

semiconductor can be used to generate electronehole pairs.

Moreover, the photocathode materials should be environ-

mentally benign, low-cost, and highly durable under cathodic

potentials and illumination.

Copper (II) oxide (CuO) is a p-type semiconductor with a

direct bandgap ranging from 1.3 eV to 1.8 eV depending on the

synthesis method and condition [10e15]. The small bandgap

energy allows CuO to absorb vast majority of solar spectrum,

and the direct bandgap endows CuO with a much larger ab-

sorption coefficient compared to the indirect bandgap mate-

rials. Therefore, CuO is believed to be a promising candidate

for water splitting photocathodes. Moreover, the low toxicity

and the relatively high abundance in earth crust make CuO

more attractive. Several works have demonstrated that CuO

can serve as a good photocathodic material by itself [13e19] or

as a component to build heterojunction photoelectrodes with

other semiconductors, such as TiO2 and ZnO [20e23].

However, as a photocathodic material, CuO has two main

shortcomings. (1) The overpotential of HER at CuO surface is

high, which is a common problem for most of the semi-

conductor oxides. (2) CuO is unstable under cathodic poten-

tials and illumination [10,15,20], and the HER photocurrent at

CuO photocathodes decays quickly with time [10,15]. It has

been proposed that CuO can be reduced during photocathodic

HER [10,20]. As we know, the reduction of CuO is a photo-

electron consuming side reaction that competes with HER,

and this reaction will inevitably lower the photo-to-hydrogen

efficiency. To solve the first issue, noble metals, such as Pt, Pd

and Rh, were deposited on the surface of semiconductor ox-

ides as a cocatalyst. The presence of noble metals can greatly

improve the HER efficiency of the semiconductors by lowering

the overpotential for HER [24]. Though this strategy has been

proved to be effective in several “semiconductor oxide/noble

metal” systems [24e29], no work has been reported on the

preparation and properties of CuO/noblemetal composites for

PEC HER. As for the second issue, we believe the strategy of

depositing noble metals on CuO surface may also be a good

solution to improve the stability of CuO because of the

following reasons. The reduction of water and the reduction of

CuO by photo-electrons are two competitive photocathodic

reactions, and the presence of cocatalyst on CuO can inhibit

the latter reaction by accelerating the former. However, no

work has been reported to confirm the feasibility of this

strategy. Most of the previous work concerning semi-

conductor/noble metal photocathodes was performed on

stable oxides, such as TiO2 and ZnO [24e29]. As a result, the

effect of noble metal cocatalysts on the stability of photo-

cathode materials has not been well addressed.

In this work, we reported the preparation of the CuO/Pd

composite photocathode by deposition of Pd nanoparticles on

the surface of CuO using a photo-assisted electrochemical

approach.Wedemonstrated that the composite photocathodes

exhibited a substantially enhanced HER photocurrent in low

overpotential region.We showed that the reduction of CuOwas

greatly restrained and the stability of photoactivity of CuO was

significantly improved after deposition of Pd nanoparticles.

Experimental section

Chemicals and materials

CuSO4$5H2O was purchased from Sinopharm Chemical Re-

agent Co., Ltd. L-ascorbic acid was purchased from the Tianjin

Jinke Fine Chemical Institute. Ethylene glycol was purchased

from Beijing Chemical Reagents Company. All chemicals were

of analytical reagent grade and used without further purifi-

cation. The fluorine-doped tin oxide (FTO) glass (8 U sq�1,

transparency 80%, Asahi Glass, Japan) was used as the

conductive substrate. All aqueous solutions were prepared

with deionized water (resistance > 18 MU cm).

Preparation of CuO and CuO/Pd photocathodes

The CuO films were obtained by thermal oxidation of Cu

nanoparticles in air. Cu nanoparticles were synthesized by

mixing the same volume of 100mMethylene glycol solution of

CuSO4 and 100 mM ethylene glycol solution of L-ascorbic acid

in room temperature, and mixture was sonicated for 5 min.

Then, the resulting Cu nanoparticles (NPs) were separated by

centrifugation at 12,000 rpm for 30 min. After being washed

with copious ethanol, the Cu NPs were dispersed in ethanol

and the final concentration was 30 mg/mL. The Cu NPs layers

were prepared on FTO substrates by spin-coating. In detail,

10 mL Cu NPs solution was dropped on FTO substrates and

then spin-coated at 1000 rpm for 15 s. The thickness of Cu NPs

film was controlled by the cycles of spin-coating, and the

sample obtained by one spin-coating cycle was referred as

one-layered Cu NPs. After spin-coating, the Cu NP modified

FTO substrates were thermal treated in a furnace in air at

required temperature (350, 450, and 550 �C) for required time.

During thermal treatment, the Cu NPs were oxidized to CuO.

The Pd NPs were deposited on the surface of CuO using a

photo-assisted electrodeposition method. In detail, the CuO

modified FTO substrates were immersed in 0.1 M KCl con-

taining 0.5 mM Na2PdCl4, and then a deposition potential of

0.1 V was applied to the CuO electrode for required time under

illumination (100 mW cm�2). The resulting CuO/Pd composite

electrodes were rinsed with copious water before they were

used in electrochemical and PEC measurements.

Electrochemical and photoelectrochemical (PEC)measurements

A three-electrode cell was used in PEC measurements. The

CuO modified FTO was employed as working electrode, a

saturated calomel electrode (SCE) and a platinum foil were

used as reference and counter electrodes, respectively. The

Fig. 1 e The SEM images of one-layered (a) and 5-layered (b) Cu NPs prepared on FTO substrates with one and five spin-

coating cycles, respectively. The SEM images of one-layered (c) and 5-layered (d) CuO films after thermal oxidation in air at

550 �C. Typical TEM images and SAED patterns of Cu NPs (e) and CuO film (f).

i n t e rn a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 3 9 ( 2 0 1 4 ) 7 6 8 6e7 6 9 67688

area of CuO electrode exposed to solution and light was

0.25 cm2. All electrochemical and PEC measurements were

carried out in 0.2 M H2BO3�/H3BO3 buffer solution (pH ¼ 9.2) on

a CHI660C electrochemical workstation (CH Instruments Co.).

All potentials were reported with respect to SCE or reversible

hydrogen electrode (RHE). Electrochemical impedance mea-

surements were performed at a frequency of 1 KHz and an

amplitude of 5 mV to determine the flat-band potential of the

CuO electrode by using MotteSchottky plot [30]. For PEC

measurements, a 500 W Xenon lamp was used as light source

and the intensity of light was 100 mW cm�2.

Characterization

The phase analysis of the Cu NPs and the CuO films were

performed by X-ray diffraction (XRD) (Rigaku, Rint 2000

advance theta-2theta powder diffractometer) with CuKa ra-

diation. The morphology and size of the Cu NPs and the CuO

films were characterized by field-emission scanning electron

microscopy (FE-SEM) (Hitachi, S-4800, Japan). High-resolution

transmission electron microscope (HRTEM) measurements

were performed on a field emission JEM-2010F microscope

(JEOL Ltd., Japan) with an accelerating voltage of 200 kV.

Selected area electron diffraction (SAED) patterns were also

recorded on the field emission JEM-2010F microscope. The

diffuse reflectance spectrumof the CuO filmwas recorded by a

double beam UVeVis spectrophotometer (Purkinje General,

China) with a bare FTO substrate as reference.

Results and discussion

Morphology, crystal structure, and bandgap of CuOelectrodes

Fig. 1(a) and (b) shows the SEM images of (one-layered and

five-layered) Cu NPs dispersed on FTO substrates by one and

five repeated spin-coating cycles, respectively. The Cu NPs

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 3 9 ( 2 0 1 4 ) 7 6 8 6e7 6 9 6 7689

synthesized by reduction of Cu2þ with L-ascorbic acid have a

quasi-spherical shape with an average diameter of about

420 nm. The FTO substrate was not completed coated with Cu

NPs after one cycle of spin-coating, as shown in Fig. 1(a).While

after 5 cycles of spin-coating, an uniform Cu NPs film was

obtained on the FTO substrates and no exposed FTO substrate

areas can be observed in SEM image (Fig. 1(b)). In fact, three

cycles of spin-coating resulted in a complete Cu NPs film that

fully coated the FTO surface. Fig. 1(b) also shows that the Cu

NPs film is porous due to the random stacking of Cu particles.

XRD measurements were carried out to determine the crystal

structure of the Cu NPs, and the result is shown in Fig. 2,

which indicates that the Cu NPs have a cubic crystal structure

(JCPDS NO. 85-1326). To obtain a CuO film, the Cu NP modified

FTO substrates were thermal treated in air at 550 �C for 4 h.

After thermal treatment, the morphology of the sample

changes from a porous film to a more continuous one, with a

significant decrease of the porosity (Fig. 1(c) and (d)). It should

be pointed out here that, though the Cu NPs layer prepared

with one cycle of spin-coating does not fully coat the FTO

surface, a continuous and complete thin layer of CuO was

formed on the FTO surface after thermal oxidation, as can be

seen by comparing Fig. 1(a) and (c). XRD spectra (Fig. 2) confirm

that the Cu NPs were oxidized to CuO during thermal treat-

ment and the resulting CuO film has a monoclinic structure

(JCPDS NO. 5-661). TEM images of Cu NPs and CuO film are

shown in Fig. 1(e) and (f), respectively. The HRTEM image and

the SAED pattern of an individual Cu NP (Fig. 1(e)) further

confirm that the Cu NPs have a single crystalline structure.

The lattice spacing shown in Fig. 1(e) is 0.204 nm, which is in

accordance with that of (111) plane of cubic Cu. This result

agrees well with the XRD spectrum of Cu NPs shown in Fig. 2.

The HRTEM image and the SAED pattern in Fig. 1(f) indicate

that the CuO film is composed of single crystalline CuO small

particles. The lattice spacing shown in Fig. 1(f) is 0.231 nm,

which matches well with that of the (111) planes of the

monoclinic CuO. It should be pointed out that lattice spacing

of 0.158 nm, which corresponds to the lattice distance of (202)

planes of monoclinic CuO, can also be observed in some

HRTEM images of CuO film (data not shown). These TEM

Fig. 2 e XRD pattern of the Cu NPs dispersed on FTO

substrate with 5 cycles of spin-coating and the

corresponding CuO film after thermal oxidation at 550 �C.

results are in good agreement with the XRD pattern of CuO

shown in Fig. 2.

TheUVevis diffuse reflection spectroscopicmeasurements

were performed to investigate the light absorption property of

the CuO films. Fig. 3(a) shows the Tauc plot, which was ob-

tained from UVevis spectra based on Tauc equation [31] and

can be used to determine the bandgap energy (Eg) of the CuO

films [32]. The bandgap energy obtained from Fig. 3(a) is

1.56 eV, suggesting that all the UV and visible light can, in

principle, be employed to generate electronehole pairs in CuO

film. Because one of the key processes for PEC HER is the in-

jection of photo-generated electrons from the conduction

band of CuO to water, the position of the conduction band (Ec)

with respect to the potential of H2O/H2 is very important for

PEC HER. According to the previous work [10], the position of

the valence band (Ev) below the vacuum level is given by the

following equation:

Ev ¼ Vfb þ 4:74þ DEF þ VH (1)

where Vfb is the flat-band potential, 4.74 eV is the position of

SCE below the vacuum level, DEF is the energy difference

Fig. 3 e (a) Tauc plot obtained from the UVevis diffuse

reflectance spectrum of the CuO film. R is the reflectance of

the CuO film and F(R) [ (1 L R)2/2R. (b) The MotteSchottky

plots of CuO film in 0.2 M H2BO3L/H3BO3 buffer solution

(pH [ 9.2). (For interpretation of the references to color in

this figure legend, the reader is referred to the web version

of this article.)

Fig. 4 e (a) Current density response of 10-layered CuO

photocathode as a function of potential in the dark and

under illumination. (b) Net photocurrent density vs

potential plot obtained on different layered CuO samples.

The different layered CuO samples were prepared by spin-

coating Cu NPs for different cycles on FTO followed by

thermal treatment at 550 �C. The current-potential curves

were obtained in 0.2 M H2BO3L/H3BO3 buffer solution

(pH [ 9.2) with a potential sweep rate of 50 mV sL1.

i n t e rn a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 3 9 ( 2 0 1 4 ) 7 6 8 6e7 6 9 67690

between Fermi level (EF) and Ev, which can be calculated by

DEF ¼ EF � Ev, and VH is the Helmholtz potential between the

CuO electrode and the solution. On the basis of previous work

[10,13], DEF is 0.13 eV for CuO. The Helmholtz potential, VH, is

given by [10]:

VH ¼ 0:059ðpHz � pHaÞ (2)

where pHz is the pH of zero charge, which is defined as the pH

value of the solution in which the surface of semiconductor is

zero charged, and for CuO$pHz is 9.5 � 0.4 [33]. pHa is the pH

value of any actual solution in which CuO is immersed. Ac-

cording to Eq. (2), when the pH value of a solution is ca. 9.5, the

Helmholtz potentialVH for CuO in this solution can be neglect.

To meet the requirement for neglecting VH, 0.2 M H2BO3�/

H3BO3 buffer solution with a pH value of 9.2 was employed as

electrolyte in this work. Then, to calculate the value of Ev from

Eq. (1), we also need to know the flat-band potential, Vfb.

In order to obtain Vfb, the space charge capacitance (C) of

CuO at CuO/solution interface was measured as a function of

applied potential (V) at a frequency of 1 kHz in 0.2 M H2BO3�/

H3BO3 buffer solution (pH ¼ 9.2), and the result is shown in

Fig. 3(b), which is also known as MotteSchottky plot. Ac-

cording to theMotteSchottky equation [30], C�2 varies linearly

with V:

1C2

¼ 2eεε0A2ND

��V � Vfb

�� kBTe

�(3)

where ε is dielectric constant of CuO, ε0 is the permittivity of

vacuum, A is the area, ND is the carrier density, e is the elec-

tron charge, kB is Boltzmann’s constant, and T is the absolute

temperature. At room temperature, kBT/e is 0.0257. Therefore,

according to MotteSchottky equation, the value of Vfb can be

obtained from the intercept of the red extrapolation line in

Fig. 3(b) [30]. The MotteShottky plot of CuO in 0.2 M H2BO3�/

H3BO3 solution exhibits a negative slope, indicating that CuO

is a p-type semiconductor, and the value of Vfb obtained from

the intercept is 0.42 V vs SCE. Therefore, according to Eq. (1), Evis calculated to be 5.29 eV, whichmeans that the valence band

of CuO lies at 5.29 eV below the vacuum level. Then, on the

basis of the bandgap energy Eg ¼ Ec � Ev ¼ 1.56 eV, the con-

duction band lies at Ec ¼ 3.73 eV below the vacuum level,

which corresponds to �1.01 V vs SCE. As the equilibrium po-

tential of the H2O/H2 redox couple lies at �0.79 V vs SCE in

0.2 M H2BO3�/H3BO3 buffer solution (pH ¼ 9.2), the conduction

band edge of CuO is more negative than the redox potential of

H2O/H2. As a result, photo-generated electrons in conduction

band can be injected into the solution to reduce water. The

above discussion confirms that the CuO prepared in this work

can be employed as photocathodematerial for water splitting.

However, it should be pointed out that the photo-generated

electrons can also reduce CuO to form Cu2O because the

conduction band edge of CuO ismuchmore negative than that

of the CuO/Cu2O redox couple (�0.06 V vs SCE at pH¼ 9.2) [10].

Therefore, the reduction of water and the reduction of CuO

are two competitive reactions with the latter reaction ther-

modynamically more favorable, though the former may be

dynamically more preferred. The reduction of CuO may show

its side effect on the HER after long reaction time. Although

the quick decay of photocurrent for HER has been reported in

the previous work on CuO photocathodes [10,15], no work

reported what happens to CuO and its effect on HER after

long-term reaction.

PEC properties of the CuO photocathodes for HER

The photoelectrochemical properties of CuO films for HER

were examined in 0.2 M H2BO3�/H3BO3 buffer solution

(pH¼ 9.2). Fig. 4(a) shows a typical current density response of

a 10-layered CuOfilm in linear potential sweepmeasurements

in the dark and under illumination (100 mW cm�2). The onset

potential for HER in the dark is ca. �0.30 V vs SCE and there is

nearly no reduction current can be observed when the po-

tential is higher than �0.30 V. However, under illumination,

the onset potential for HER is positively shifted to 0.1 V vs SCE,

and a large reduction photocurrent is obtained even when the

potential is higher than �0.30 V. These results indicate that

Fig. 5 e (a) Current density-potential responses of the 10-

layered CuO films under chopped illumination

(100 mW cmL2). The 10-layered CuO samples were

prepared under different thermal oxidation temperatures.

(b) Current density-potential responses of the 10-layered

CuO samples in the dark and under illumination. The CuO

samples were thermal treated at 550 �C for different time.

All linear potential sweep measurements were performed

in 0.2 M H2BO3L/H3BO3 buffer solution (pH [ 9.2) with a

potential sweep rate of 50 mV sL1.

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 3 9 ( 2 0 1 4 ) 7 6 8 6e7 6 9 6 7691

the photo-generated electrons in CuO are injected into solu-

tion and contribute greatly to HER. As can be seen in Fig. 4(a),

there is a large reduction current in the dark when the po-

tential is lower than �0.30 V. To extract the pure effect of

illumination on the reduction current, the net photocurrent

(Jph) was calculated out by deducting the current in the dark

from that under illumination. Fig. 4(b) shows the variation of

Jph as a function of the applied potential and CuO layers. For all

the CuO photocathodes, Jph first increases with decreasing the

applied potentials, and then reaches its maximum at ca.

�0.50 V, and still then decreases as the potential is further

negatively shifted. Similar phenomena were observed previ-

ously on the CuO film prepared by flame spray pyrolysis [13]. It

is understandable for Jph to increasewith decreasing potential,

because lowering the applied potential will inevitably in-

creases the driving force for photo-generated electrons to

transfer from CuO to water. As for the decrease of Jph when

potential is more negative than �0.50 V, we believe it is due to

the partial reduction of CuO at very low potentials. Fig. 4(b)

also demonstrates that the number of CuO layers has signifi-

cant effect on Jph, and the maximum value of Jph is obtained

from the 10-layered CuO film. The increase of CuO thickness

has two opposite effects on the photocurrent. On the one

hand, it results in the increase of light absorption, which will

benefit Jph. While on the other hand, it also leads to the in-

crease of resistance, which will surely lower Jph.

The effects of the thermal oxidation temperature and

duration on the photoactivity of CuO toward HER were also

investigated. Three CuO photocathodes prepared under

different thermal oxidation temperature were used to record

the current-potential curve under chopped illumination. The

results are shown in Fig. 5(a), fromwhich it is seen that, raising

the oxidation temperature enhances the photocurrent, and the

highest photocurrent is obtained with a thermal treatment

temperature of 550 �C. We believe this is due to the improve-

ment of crystallinity with increasing temperature. Low tem-

perature thermal treatment results in poor crystalline quality

and then leads to a high density of defects, whichwill act as the

recombination centers for the photo-generated electronehole

pairs [31,34]. It should be pointed out here that raising the

temperature over 550 �Cwill deteriorate the conductivity of the

FTO substrate, and then significantly lower the photocurrent.

Fig. 5(b) shows the influence of thermal treatment duration on

the photoactivity of CuO films. The photocurrent density first

increases and then decreases with increasing oxidation time.

As we know, increasing thermal treatment time at 550 �Ccauses two opposite effects on the photocurrent. First, it im-

proves the crystallinity of CuO and reduces the possibility of

electronehole recombination, and this effect favors the

enhancement of photocurrent. Second, long-time thermal

treatment at 550 �C worsens the electrical properties of FTO

substrate, and then decreases the photocurrent. Accordingly,

there exists an optimal thermal oxidation time for CuO pho-

tocathodes. In consistence with the above discussion, the CuO

sample prepared by thermal oxidation at 550 �C for 4 h exhibits

the highest photoactivity, as shown in Fig. 5(b).

According to the above results, 10-layered CuO samples,

which were thermal treated in air at 550 �C for 4 h, were

employed to study the stability of CuO photocathodes. When

the CuO photocathode was repeatedly scanned within the

potential region from 0.10 V to �0.60 V under continuous or

chopped illumination, a clear decrease of photocurrent with

increasing potential sweep cycles was observed, indicating a

significant decay of the photoactivity. Moreover, the lower the

negative potential sweep end was, the faster the decay of

photoactivity. XRD spectra demonstrate that part of the sur-

face CuO was reduced to Cu2O, providing solid evidence that

the reduction of CuO is not negligible for long-term PEC HER.

As the reduction of CuO by photo-induced electrons is a par-

allel competitive reaction with HER, lowering the over-

potential for HER not only enhances the photoactivity for HER

but also inhibits the reduction of CuO. As Pd has a very low

overpotential for HER [24], Pd nanoparticles (Pd NPs) were

deposited on CuO to prepare CuO/Pd composite, and then

to improve the efficiency and stability of CuO-based

photocathodes.

Fig. 6 e Typical SEM (a) and HRTEM (b) images of Pd/CuO composite prepared by photo-assisted electrodeposition at 0.10 V

for 15 min in 0.1 M KCl containing 0.5 mM Na2PdCl4. The inset in (a) is a high-resolution SEM image.

i n t e rn a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 3 9 ( 2 0 1 4 ) 7 6 8 6e7 6 9 67692

PEC properties of the CuO/Pd composite photocathodes forHER

The CuO/Pd composites have been employed as catalysts in

several reactions, including Suzuki reaction [35], cyanation

reaction [35], hydrogen generation reaction by methanol par-

tial oxidation [36], and photodegradation [37]. In this work, the

Pd is used as an electrocatalyst, and the modification of CuO

photocathode with Pd is aimed facilitating HER by lowering

the HER overpotential. The CuO/Pd composite photocathodes

were prepared by photo-assisted electrodeposition of Pd on

the surface of CuO film. Under illumination, the photo-

generated electrons can reduce the high-valence palladium

precursor (PdCl42�) to zero-valence metal Pd NPs. As a result,

the produced Pd NPs can be loaded on the surface of CuO film.

The applied potential is much lower than the redox potential

of PdCl42�/Pd couple, and then can facilitate the reduction of

PdCl42� by photo-induced electrons. The advantage of this

deposition method is that PdCl42� is directly reduced to Pd at

the photoactive sites on CuO, which ensures a faster electron

transfer fromCuO to Pd. Themorphology of the resulting CuO/

Pd composite is shown in Fig. 6(a), fromwhich a uniform layer

of small nanoparticles with particle size ranging from 3 nm to

20 nm is observed on the surface of CuO. Energy dispersive

spectroscopy (EDS) was used to identify the elemental com-

positions of the particle modified CuO. Pd peaks appear on

EDS spectra (see Fig. S1), providing the solid evidence that the

small particles in SEM image (Fig. 6) are Pd NPs. Moreover, the

compositional analysis by EDS indicates that, for the CuO/Pd

composite prepared by photo-assisted electrodeposition of Pd

for 15min, the atomic weight percentage of Pd on CuO surface

is 3.57%. A typical HRTEM image of CuO/Pd composite is

shown in Fig. 6(b), from which the lattice fringes of both CuO

and Pd are clearly seen. The Pd particles have a single crys-

talline structure with a lattice spacing of 0.225 nm, in accor-

dance with that of (111) facet of Pd. The CuO film exhibits a

lattice spacing of 0.158 nm, which corresponds to the lattice

distance of (202) facet of monoclinic CuO. All these SEM, EDS

and TEM results clearly demonstrate that small Pd nano-

crystals can be successfully generated on the surface of CuO

by photo-assisted electrodeposition.

As shown in Figs. 4(a) and 5, a relatively large dark reduc-

tion current is observed on CuO photocathodes when the

potential is more negative than 0.35 V. This current may

partially come from the electrochemical reduction of CuO.

Accordingly, to avoid electrochemical reduction of CuO, PEC

measurements on CuO/Pd composite photocathodes were

performed with the applied potential higher than �0.35 V.

Fig. 7(a) demonstrates the current-potential responses of CuO

film in the dark and under illumination before and after Pd

deposition. Because of the low HER overpotential of the Pd

cocatalyst [24], the CuO/Pd composite exhibits a much larger

reduction current under illumination compared to CuO. In

fact, after Pd deposition, the photocurrents increase by ca.

110%, 105%, 59%, 21% at potentials of 0.0, �0.10, �0.20 and

�0.30 V, respectively, indicating an enhancement of photo-

activity especially at low overpotentials. Electrochemical

impedance spectroscopy (EIS) was employed to provide the

information of the electron transfer across CuO/solution

interface. Fig. 7(b) shows the Nyquist plots of the CuO and the

CuO/Pd photocathodes at a constant potential of �0.35 V.

Clearly there are two semicircles apparent in the EIS spectra

on both photocathodes. The semicircle in high frequency re-

gion is attributed to the impedance of the CuO/FTO interface,

which can be represented by the CuO/FTO contact resistance

and the CuO/FTO interfacial capacitance in parallel. This

semicircle keeps almost unchanged regardless of illumination

and the presence of Pd because the contact resistance and the

capacitance of CuO/FTO are independent of illumination and

Pd modification. The low frequency semicircle corresponds to

the impedance of the CuO/solution interface. The diameter of

this semicircle represents the charge transfer resistance (Rct)

across the CuO/solution interface. Fig. 7(b) clearly shows that,

for the CuO/Pd electrode, Rct under illumination is much

smaller than that in the dark, indicating that photo-induced

electrons are involved in the electron transfer across the

CuO/solution interface. Moreover, under continuous illumi-

nation, the value of Rct (108 U cm2) on the CuO/Pd electrode is

much smaller than that on the CuO electrode (258 U cm2),

providing another evidence that the deposition of Pd on CuO

surface facilitates the transfer of photo-induced electrons

from CuO to solution.

The amount of deposited Pd, which can be adjusted by

varying the deposition time, is a key factor controlling the

photoactivity of CuO/Pd composite. Fig. 8 shows the influence

of Pd deposition time on the photocurrent of CuO/Pd

Fig. 7 e (a) Current density-potential curves of the CuO

films in the dark and under illumination before and after

photo-assisted electrodeposition of Pd for 15 min. The

potential sweep rate is 50mV sL1. (b) EIS spectra of the CuO

and the Pd/CuO photocathodes at an applied potential of

L0.35 V. All the electrochemical measurements were

performed in 0.2 M H2BO3L/H3BO3 buffer solution (pH[ 9.2).

The Pd/CuO composite photocathodes were prepared by

photo-assisted deposition of Pd on CuO at 0.1 V for 15 min

in 0.1 M KCl containing 0.5 mM Na2PdCl4.

Fig. 8 e Current density-potential responses of the Pd/CuO

composite photocathodes with different Pd deposition

time. The Pd/CuO composite photocathodes were prepared

by photo-assisted deposition of Pd on CuO at 0.1 V for

different time in 0.1 M KCl containing 0.5 mM Na2PdCl4.

The linear potential sweep measurements were performed

in 0.2 M H2BO3L/H3BO3 buffer solution (pH [ 9.2) with a

potential sweep rate of 50 mV sL1 in the dark and under

illumination.

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 3 9 ( 2 0 1 4 ) 7 6 8 6e7 6 9 6 7693

composites. The photocurrent first increases and then de-

creases with increasing deposition time, and a maximum

photocurrent was obtained with a deposition time of 15 min.

This result is due to the fact that the amount of deposited Pd

influences the photoactivity of CuO/Pd in two opposite di-

rections. A large Pd deposition amount (long deposition time)

can surely benefit the electron transfer from CuO to water.

However, it may also inhibit the light absorption of CuO and

then lowers the photoactivity. We believe that the maximum

photocurrent shown in Fig. 8 is a balance between the two

opposite effects.

The photocatalytic stability of the CuO and the CuO/Pd

photocathodes were examined by using chronoamperometry.

Fig. 9(a) show the variation of photocurrent density of CuO

and CuO/Pd as a function of time under chopped illumination

at fixed potentials of �0.35 V. According to the discussion

in PEC properties of the CuO photocathodes for HER, the

potential of �0.35 V was used in the chronoamperometric

measurements because it is the most negative potential that

can produce the largest HER current while avoiding the elec-

trochemical reduction of CuO. Fig. 9(a) clearly demonstrates

that, the photocurrent of CuO decreases to approximately 19%

of its original value after 10 h reaction, whereas the photo-

current of the CuO/Pd composite decays more slowly

compared to that of CuO. After 10 h reaction, the remaining

photocurrent density of the composite photocathode is ca.

0.30 mA cm�2, which is over two times larger than that of the

CuO electrode. In addition, after reaction for the first 3 h, the

photocurrent on CuO/Pd keeps almost constant, only less than

15% photocurrent decay is observed for the next 7 h. However,

as for the CuO film, the photocurrent decreases by over 60%

for the last 7 h. These results provide solid evidence that the

combination of Pd with CuO can significantly improve the

photocatalytic stability of the photocathode. Fig. 9(a) also re-

veals that, even on the CuO/Pd composite photocathode, the

photocurrent density still decreases with increasing reaction

time. This is due to the fact that the CuO surface was not

completely coated with Pd NPs, as can be seen in the SEM

image in Fig. 6. Therefore, the CuO surface exposed to solution

on the CuO/Pd electrode may be reduced by photo-induced

electrons and then leads to the decrease of photocurrent.

When the exposed active sites on CuO surface were reduced

by photo-induced electrons and lost their activity, the rate of

current decay on the CuO/Pd composite slowed down, as can

be seen from the stable photocurrent for the last seven hours

(Fig. 9(a)).

It should be pointed out here that, even at the same po-

tentials, the photocurrents obtained from chronoampero-

metric measurements under chopped illumination on both

the CuO and the CuO/Pd photocathodes (Fig. 9(a)) are smaller

Fig. 9 e (a) Current density response as a function reaction

time on the CuO and CuO/Pd photocathodes at a constant

applied potential of L0.35 V under chopped illumination.

(b) XRD spectra of the CuO and the Pd/CuO photocathodes

after held at L0.35 V for 10 h under chopped illumination.

Fig. 10 e The band energy diagrams of the CuO/Cu2O (a)

and the Pd/CuO (b) photocathodes in 0.2 M H2BO3L/H3BO3

buffer solution (pH [ 9.2).

i n t e rn a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 3 9 ( 2 0 1 4 ) 7 6 8 6e7 6 9 67694

than those in linear potential sweep measurements under

continuous illumination (Fig. 7(a)).We believe the difference is

due to the different methodology used to obtain photocur-

rents, and the detailed reason is not fully understood.

After long-time reaction under illumination, the phase

structure of the CuO and the CuO/Pd electrodeswere analyzed

by XRD. Fig. 9(b) shows the XRD spectra of the CuO and the

CuO/Pd electrodes after they were held at �0.35 V for 10 h

under chopped illumination. For both electrodes, two new

diffraction peaks at 36.4� and 42.4� appear besides the peaks of

monoclinic CuO. These two peaks correspond to the (111) and

(200) diffraction peaks of cubic Cu2O, as can be seen from

standard XRD pattern of cubic Cu2O (JCPDS No. 77-199). No

diffraction peaks from metal Cu can be observed, indicating

that no detectable amount of Cuwas generated as the product

of the two-electron reduction of CuO. These XRD results pro-

vide direct and solid evidence that part of CuO is reduced by

photo-generated electrons at �0.35 V under illumination, and

the main product is Cu2O. The intensity ratio of the (111)

diffraction peak of Cu2O to the ð111Þ diffraction peak of CuO,

which were denoted as g, reflects the ratio of the amount of

Cu2O to that of CuO. In other words, the value of g indicates, to

some extent, the percentage of the surface CuO that was

reduced to Cu2O. As shown in Fig. 9(b), after ten hour’s reac-

tion, the g values for the CuO and the CuO/Pd electrodes are

0.77 and 0.35, respectively, with the latter less than half of the

former. This result reveals the fact that the amount of surface

CuO, which were reduced to Cu2O on the CuO/Pd composite

photocathode during PEC HER, is much smaller than that on

the CuO photocathode. Therefore, it can be safely concluded

that the deposition of Pd cocatalyst not only facilitates the

electron transfer across CuO/solution interface, but also

significantly inhibits the reduction of CuO to Cu2O.

As is well known, Cu2O is a p-type semiconductor with a

direct bandgap energy of ca. 2.0 eV, which is also suitable for

solar energy harvesting [7]. The conduction band position of

Cu2O is located at 0.7 V negative of the hydrogen evolution

potential, making Cu2O a suitable photocathode material for

HER [7,38]. Then, an important issue arises: why the presence

of Cu2O on the surface of CuO lowers photocurrentwhen Cu2O

is also a photocathode material? We believe this is due to the

mismatch in the band structure between CuO and Cu2O.

Fig. 10(a) shows the band energy diagrams of the CuO/Cu2O/

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 3 9 ( 2 0 1 4 ) 7 6 8 6e7 6 9 6 7695

solution heterojunction, which are drawn based on the

bandgap energy and the conduction band position of CuO and

Cu2O. Since Cu2O is generated on the surface of CuO, photo-

induced electrons have to pass through the CuO/Cu2O/solu-

tion interface to reducewater. However, as shown in Fig. 10(a),

the conduction band position of Cu2O is higher (more nega-

tive) than that of CuO, making it impossible for the photo-

induced electrons to transfer from CuO to Cu2O [39]. More-

over, some of the photo-generated electrons at Cu2O can be

injected into CuO, and accumulate in the conduction band of

CuO. This process greatly increases the possibility of electro-

nehole recombination inside CuO and decreases the

photocurrent.

Fig. 10(b) shows the band energy diagram of the CuO/Pd

composite in solution. Because the conduction band position

of CuO is 0.22 V negative of hydrogen evolution potential in

0.2 M H2BO3�/H3BO3 buffer solution (pH ¼ 9.2), the photo-

generated electrons have two alternative paths to transfer

from CuO to water on CuO/Pd electrode. (1) Electrons transfer

directly fromCuO towater, and (2) electrons first transfer from

CuO to Pd and then from Pd towater, with Pd as amediator. As

is well known, Pd exhibits a high electrochemical HER activity

[40], which ensures a very low HER overpotential on Pd [24].

Therefore, the latter path via Pd mediator has a much faster

electron transfer rate than the former one. This is the reason

why Pd deposition greatly improves the photocurrent for HER.

As we have demonstrated before, two competitive reactions

are involved during PEC HER: the reduction of water to

generate H2 and the reduction of CuO to form Cu2O. The

presence of Pd on the surface of CuO can greatly accelerate the

electron transfer fromCuO towater, andwill surely inhibit the

reduction of CuO by photo-generated electrons. That’s why

the deposition of Pd significantly improves the PEC stability of

the CuO photocathodes.

It should be pointed out here that the purpose of this work

is to demonstrate that the fabrication of CuO/Pd composite

photocathodes by depositing Pd on the surface of CuO can

greatly enhance the activity and stability of the photocathode

toward PEC HER. We do not take much effort to improve the

photocurrent on CuO film. Therefore, this leaves vast spaces

for further improving the photoactivity of the CuO/Pd com-

posite electrodes by, for example, increasing the specific area

(porosity) of the CuO film.

Conclusion

The monoclinic CuO films with bandgap energy of 1.56 eV

were prepared by a facile and cost-effective method, which

included solution reduction, spin-coating and thermal

oxidation processes. The conduction band edge of the CuO

film was located �1.01 V vs SCE in aqueous solution with

pH ¼ 9.2, making CuO a good candidate for water splitting

photocathode. The highest photoactivity toward HER was

obtained on 10-layered CuO films that were thermal treated at

550 �C for 4 h. We provided the first solid evidence that part of

the surface CuO was reduced to Cu2O by photo-induced

electrons during PEC HER. The reduction of CuO greatly low-

ered the photoactivity and photostability of the CuO film. A

photo-assisted electrodeposition method that ensured the

deposition of Pd on the photoactive sites of CuO surface was

developed to prepare CuO/Pd composite photocathodes. We

demonstrated that the deposition of Pd on CuO not only

enhanced the photocurrent for HER but also significantly

improved the photocatalytic stability of the CuO film. This

work shows the feasibility of increasing PEC activity and sta-

bility of CuO-based photocathodes by developing CuO/noble

metal composites.

Acknowledgments

The authors acknowledge the financial support of thiswork by

National Natural Science Foundation of China (NSFC

21173016, 20973020), Beijing Natural Science Foundation

(2142020), Doctoral Fund of Ministry of Education of China

(20101102110002).

Appendix A. Supplementary data

Supplementary data related to this article can be found at

http://dx.doi.org/10.1016/j.ijhydene.2014.03.084.

r e f e r e n c e s

[1] Lewis NS, Nocera DG. Powering the planet: chemicalchallenges in solar energy utilization. Proc Natl Acad Sci USA2006;103:15729e35.

[2] Fujishima A, Honda K. Electrochemical photolysis of water ata semiconductor electrode. Nature 1972;238:37e8.

[3] Gratzel M. Photoelectrochemical cells. Nature2001;414:338e44.

[4] Walter MG, Warren EL, McKone JR, Boettcher SW, Mi Q,Santori EA, et al. Solar water splitting cells. Chem Rev2010;110:6446e73.

[5] Li ZS, Luo WJ, Zhang ML, Feng JY, Zou ZG.Photoelectrochemical cells for solar hydrogen production:current state of promising photoelectrodes, methods toimprove their properties, and outlook. Energy Environ Sci2013;6:347e70.

[6] Hou Y, Abrams BL, Vesborg PCK, Bjorketun ME, Herbst K,Bech L, et al. Bioinspired molecular co-catalysts bonded to asilicon photocathode for solar hydrogen evolution. Nat Mater2011;10:434e8.

[7] Paracchino A, Laporte V, Sivula K, Gratzel M, Thimsen E.Highly active oxide photocathode for photoelectrochemicalwater reduction. Nat Mater 2011;10:456e61.

[8] McKone JR, Pieterick AP, Gray HB, Lewis NS. Hydrogenevolution from Pt/Ru-coated p-type WSe2 photocathodes. JAm Chem Soc 2013;135:223e31.

[9] Bak T, Nowotny J, Rekas M, Sorrell CC. Photo-electrochemicalhydrogen generation from water using solar energy.Materials-related aspects. Int J Hydrogen Energy2002;27:991e1022.

[10] Koffyberg FP. A photoelectrochemical determination of theposition of the conduction and valence band edges of p-typeCuO. J Appl Phys 1982;53:1173.

[11] Oral AY, Mens‚ur E, Aslan MH, Bas‚aran E. The preparation ofcopper(II) oxide thin films and the study of theirmicrostructures and optical properties. Mater Chem Phys2004;83:140e4.

i n t e rn a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 3 9 ( 2 0 1 4 ) 7 6 8 6e7 6 9 67696

[12] Nakaoka K, Ueyama J, Ogura K. Photoelectrochemicalbehavior of electrodeposited CuO and Cu2O thin films onconducting substrates. J Electrochem Soc 2004;151:C661e5.

[13] Chiang C-Y, Aroh K, Franson N, Satsangi VR, Dass S,Ehrman S. Copper oxide nanoparticle made by flame spraypyrolysis for photoelectrochemical water splitting e part II.Photoelectrochemical study. Int J Hydrogen Energy2011;36:15519e26.

[14] Izaki M, Nagai M, Maeda K, Mohamad FB, Motomura K,Sasano J, et al. Electrodeposition of 1.4-eV-bandgap p-copper(II) oxide film with excellent photoactivity. J Electrochem Soc2011;158:D578e84.

[15] Chiang C-Y, Chang M-H, Liu H-S, Tai CY, Ehrman S. Processintensification in the production of photocatalysts for solarhydrogen generation. Ind Eng Chem Res 2012;51:5207e15.

[16] Chauhan D, Satsangi VR, Dass S, Shrivastav R. Preparationand characterization of nanostructured CuO thin films forphotoelectrochemical splitting of water. Bull Mat Sci2006;29:709e16.

[17] Chiang C-Y, Epstein J, Brown A, Munday JN, Culver JN,Ehrman S. Biological templates for antireflective currentcollectors for photoelectrochemical cell applications. NanoLett 2012;12:6005e11.

[18] Chiang C-Y, Shin Y, Aroh K, Ehrman S. Copper oxidephotocathodes prepared by a solution based process. Int JHydrogen Energy 2012;37:8232e9.

[19] Barreca D, Fornasiero P, Gasparotto A, Gombac V, Maccato C,Montini T, et al. The potential of supported Cu2O and CuOnanosystems in photocatalytic H2 production.ChemSusChem 2009;2:230e3.

[20] Choi H-J, Kang M. Hydrogen production from methanol/waterdecomposition in a liquid photosystem using the anatasestructure of Cu loaded. Int J Hydrogen Energy 2007;32:3841e8.

[21] Liu Z, Bai H, Xu S, Sun DD. Hierarchical CuO/ZnO “corn-like”architecture for photocatalytic hydrogen generation. Int JHydrogen Energy 2011;36:13473e80.

[22] Jin Z, Zhang X, Li Y, Li S, Lu G. 5.1% apparent quantumefficiency for stable hydrogen generation over eosin-sensitized CuO/TiO2 photocatalyst under visible lightirradiation. Catal Commun 2007;8:1267e73.

[23] Zhao X, Wang P, Li B. CuO/ZnO core/shell heterostructurenanowire arrays: synthesis, optical property, and energyapplication. Chem Commun 2010;46:6768.

[24] Leung DYC, Fu X, Wang C, Ni M, Leung MKH, Wang X, et al.Hydrogen production over titania-based photocatalysts.ChemSusChem 2010;3:681e94.

[25] Zou JJ, He H, Cui L, Du HY. Highly efficient Pt/TiO2

photocatalyst for hydrogen generation prepared by a coldplasma method. Int J Hydrogen Energy 2007;32:1762e70.

[26] Antony RP, Mathews T, Ramesh C, Murugesan N, Dasgupta A,Dhara S, et al. Efficient photocatalytic hydrogen generation byPt modified TiO2 nanotubes fabricated by rapid breakdownanodization. Int J Hydrogen Energy 2012;37:8268e76.

[27] Navarro RM, del Valle F, Fierro JLG. Photocatalytic hydrogenevolution from CdSeZnOeCdO systems under visible lightirradiation: effect of thermal treatment and presence of Ptand Ru cocatalysts. Int J Hydrogen Energy 2008;33:4265e73.

[28] Rosseler O, Shankar MV, Du MKL, Schmidlin L, Keller N,Keller V. Solar light photocatalytic hydrogen production fromwater over Pt and Au/TiO2(anatase/rutile) photocatalysts:influence of noble metal and porogen promotion. J Catal2010;269:179e90.

[29] Sreethawong T, Yoshikawa S. Comparative investigation onphotocatalytic hydrogen evolution over Cu-, Pd-, and Au-loaded mesoporous TiO2 photocatalysts. Catal Commun2005;6:661e8.

[30] Bard AJ, Faulkner LR. Electrochemical methods:fundamentals and applications. 2nd ed. New York: JohnWiley & Sons, Inc; 2001.

[31] Tauc J, Grigorovici R, Vancu A. Optical properties andelectronic structure of amorphous germanium. Phys StatusSolidi B 1966;15:627e37.

[32] Sunkara S, Vendra VK, Kim JH, Druffel T, Sunkara MK.Scalable synthesis and photoelectrochemical properties ofcopper oxide nanowire arrays and films. Catal Today2013;199:27e35.

[33] Parks GA. The isoelectric points of solid oxides, solidhydroxides, and aqueous hydroxo complex systems. ChemRev 1965;65:177e98.

[34] Navarro Yerga RM, Alvarez Galvan MC, del Valle F, Villoria dela Mano JA, Fierro JLG. Water splitting on semiconductorcatalysts under visible-light irradiation. ChemSusChem2009;2:471e85.

[35] Chattopadhyay K, Dey R, Ranu BC. Shape-dependentcatalytic activity of copper oxide-supported Pd(0)nanoparticles for Suzuki and cyanation reactions.Tetrahedron Lett 2009;50:3164e7.

[36] Schuyten S, Guerrero S, Miller JT, Shibata T, Wolf EE.Characterization and oxidation states of Cu and Pd inPdeCuO/ZnO/ZrO2 catalysts for hydrogen production bymethanol partial oxidation. Appl Catal a e Gen2009;352:133e44.

[37] Huo C, Yang H. Preparation and enhanced photocatalyticactivity of PdeCuO/palygorskite nanocomposites. Appl ClaySci 2013;74:87e94.

[38] de Jongh PE, Vanmaekelbergh D, Kelly JJ.Photoelectrochemistry of electrodeposited Cu2O. JElectrochem Soc 2000;147:486e9.

[39] Zhang Z, Wang P. Highly stable copper oxide composite as aneffective photocathode for water splitting via a facileelectrochemical synthesis strategy. J Mater Chem2012;22:2456.

[40] Grigoriev SA, Millet P, Fateev VN. Evaluation of carbon-supported Pt and Pd nanoparticles for the hydrogenevolution reaction in PEM water electrolysers. J PowerSources 2008;177:281e5.