ww.sciencedirect.com

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 6 ( 2 0 1 1 ) 1 5 5 2 7e1 5 5 3 7

Available online at w

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

Plasma-assisted synthesis of Ag/ZnO nanocomposites: Firstexample of photo-induced H2 production and sensing

Quentin Simon a, Davide Barreca b,*, Daniela Bekermann a, Alberto Gasparotto a,Chiara Maccato a, Elisabetta Comini c, Valentina Gombac d, Paolo Fornasiero d,Oleg I. Lebedev e, Stuart Turner f, Anjana Devi g, Roland A. Fischer g, Gustaaf Van Tendeloo f

aDepartment of Chemistry, Padova University and INSTM, 35131 Padova, ItalybCNR-ISTM and INSTM, Department of Chemistry, Padova University, 35131 Padova, ItalycCNR-IDASC, SENSOR Lab, Department of Chemistry and Physics, Brescia University, 25133 Brescia, ItalydDepartment of Chemical and Pharmaceutical Sciences, Trieste University, INSTM and ICCOM Trieste Research Unit, 34127 Trieste, Italye Laboratoire CRISMAT, UMR 6508, CNRS-ENSICAEN, 14050 Caen CEDEX 4, FrancefEMAT, Antwerp University, 2020 Antwerpen, Belgiumg Lehrstuhl fur Anorganische Chemie II, Ruhr-University Bochum, 44780 Bochum, Germany

a r t i c l e i n f o

Article history:

Received 26 July 2011

Received in revised form

1 September 2011

Accepted 12 September 2011

Available online 6 October 2011

Keywords:

ZnO

Ag

PE-CVD

RF-Sputtering

Hydrogen production

Hydrogen sensing

* Corresponding author. Tel.: þ39 (0) 4982751E-mail address: davide.barreca@unipd.it

0360-3199/$ e see front matter Copyright ªdoi:10.1016/j.ijhydene.2011.09.045

a b s t r a c t

Ag/ZnO nanocomposites were developed by a plasma-assisted approach. The adopted

strategy exploits the advantages of Plasma Enhanced-Chemical Vapor Deposition (PE-CVD)

for the growth of columnar ZnO arrays on Si(100) and Al2O3 substrates, in synergy with the

infiltration power of the Radio Frequency (RF)-sputtering technique for the subsequent

dispersion of different amounts of Ag nanoparticles (NPs). The resulting composites, both

as-prepared and after annealing in air, were thoroughly characterized with particular

attention on their morphological organization, structure and composition. For the first

time, the above systems have been used as catalysts in the production of hydrogen by

photo-reforming of alcoholic solutions, yielding a stable H2 evolution even by the sole use

of simulated solar radiation. In addition, Ag/ZnO nanocomposites presented an excellent

response in the gas-phase detection of H2, opening attractive perspectives for advanced

technological applications.

Copyright ª 2011, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights

reserved.

1. Introduction economy [6]. To this regard, the development of materials

The importance of hydrogen as an appealing energy vector,

due to its high efficiency and environment-friendly combus-

tion, is nowadays well recognized and documented [1e5].

Nevertheless, in spite of several research activities in this

field, the large-scale production and detection of H2, an

odorless, colorless and flammable gas, are still challenging

issues in view of the eventual transition to an H2-based

70; fax: þ39 (0) 498275161(D. Barreca).2011, Hydrogen Energy P

capable of acting as multi-functional platforms for the

sustainable generation and sensing of hydrogen, though rep-

resenting a strategic target, is still far from being completely

satisfied [7,8].

Among the possible candidates, ZnO, a n-type semi-

conductor (Eg ¼ 3.4 eV) that can be synthesized with diverse

morphologies [9], has been the object of several studies

regarding photocatalysis and gas sensing utilization [10]. The

.

ublications, LLC. Published by Elsevier Ltd. All rights reserved.

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 6 ( 2 0 1 1 ) 1 5 5 2 7e1 5 5 3 715528

functional properties for such applications can be improved

by the synergistic combination of nanostructured ZnO with

metallic NPs, which, beside their inherent catalytic activity,

promote an enhancement of charge carrier separation due to

the formation of a Schottky barrier at the metal/oxide inter-

face [11e13]. Among the different metals, silver has demon-

strated a great potential to improve ZnO antibacterial/

pollutant degradation activity [14e17] and optical properties

[18e22], whereas H2 production and sensing from Ag/ZnO

nanomaterials are still rather unexplored. In particular,

whereas few reports devoted to H2 production by methanol

reforming over Ag/ZnO catalysts are available [23,24], photo-

induced reforming of aqueous solutions to yield H2, a prom-

ising route for its clean and sustainable generation [25], has

never been attempted on such systems. In addition, only one

paper on hydrogen sensing by Ag/ZnO composites is available

in the literature [26].

Since Ag/ZnO features are strongly dependent on size,

shape, and mutual distribution of the components, the

development of controllable synthetic strategies is highly

demanded to tune their functional performances. Up to date,

Ag/ZnO nanosystems have been prepared both as powders

and supported materials by liquid [17,20], vapor [16,21], and

hybrid liquidevapor phase routes [22]. Nevertheless, in spite

of various studies, a thorough understanding of Ag/ZnO

interactions is still an open question, especially due to the

high reactivity of silver NPs towards oxidation [27,28].

Furthermore, beyond the use of powdered materials [17,29],

a current challenge concerns the development of organized

Ag/ZnO assemblies onto suitable substrates, enabling their

direct integration into functional devices.

Basing on our previous studies on single-component ZnO

and Ag nanosystems [30,31], the goal of this paper is to

propose an innovative plasma-assisted route to synthesize

supported Ag/ZnO nanocomposites with tailored features.

The adopted approach consists of two subsequent steps: i)

growth of columnar ZnO arrays by PE-CVD; ii) deposition of Ag

NPs by RF-sputtering, tuning the overall metal amount by

variation of the sputtering time. A key point of the present

strategy is the use of relatively mild processing conditions,

taking advantage of the unique activation mechanisms

enabled by cold plasmas [32]. Si(100) and polycrystalline Al2O3

have been adopted as substrates in view of H2 generation and

gas sensing tests, respectively. For the first time, the excellent

performances of the obtained materials in H2 production by

photo-reforming, as well as in H2 sensing, are presented and

discussed in relation to their chemical and physical

properties.

Table 1 e Growth conditions and silver particle sizes foras-prepared (A) and annealed (T) Ag/ZnOnanocomposites. Surface silver molar fraction,determined from XPS analyses, is defined as xAg [ [Ag/(Ag D Zn D O)]3100.

Ag sputteringtime (min)

Ag particle size(as-prepared, A)

(nm)

Ag particlesize (thermallytreated, T) (nm)

xAg(A)

xAg(T)

30 10 10 15.5 1.6

90 35 80 32.6 8.3

150 60 100 44.6 14.7

2. Experimental

2.1. Synthesis and characterization

Columnar ZnO arrays were grown by PE-CVD using a Zn(II)

bis(ketoiminate)precursorZn[(R)NC(CH3)¼C(H)C(CH3)¼O]2with

R¼ e(CH2)3OCH3 [33]. Depositions were performed onto Si(100)

substrates (MEMC�, Merano, Italy; dimensions ¼ 2 cm � 1

cm� 1mm), forhydrogenproduction, andpolycrystallineAl2O3

(dimensions ¼ 3 mm � 3 mm � 250 mm), for sensing

applications. The substrates were suitably pre-cleaned using

well established procedures [30,34].

The precursor, placed in an external vessel heated by an oil

bath, was vaporised at 150 �C and transported into the reac-

tion chamber by means of an Ar flow (60 sccm). The temper-

ature of the lines was maintained at 170 �C in order to avoid

condensation phenomena. Additional Ar (15 sccm) and O2

(20 sccm) flows were introduced into the reactor. Depositions

were performed for 1 h at 300 �C, at a total pressure of 1 mbar,

and using an RF-power of 20 W. Relevant data on the char-

acterization of ZnO matrices are provided in the Supplemen-

tary Content.

Subsequently, Ag NPs were deposited on the obtained ZnO

arrays by RF-sputtering from Ar plasmas, using an RF-power

of 5 W and a total pressure of 0.3 mbar. The process was

performed at a temperature as low as 60 �C to prevent unde-

sired modifications of the pristine ZnO matrices. Three

different sputtering times (30, 90, and 150 min, see Table 1)

were used to vary the overall silver amount in the resulting

composites.

The effect of ex-situ thermal treatment in air at 400 �C for

1 h was also investigated. In the following, samples are

denoted as X_Y, where X is the sputtering time (in min) and Y

indicates as-prepared (A) or thermally treated (T) samples,

respectively.

Field emission-scanning electronmicroscopy (FE-SEM) and

energy dispersive X-ray spectroscopy (EDXS) measurements

were carried out by a Zeiss SUPRA 40VP instrument, equipped

with an Oxford INCA x-sight X-ray detector, at acceleration

voltages between 10 and 20 kV. For cross-section EDXS line-

scans, OKa1, ZnKa1, and AgLa1 signals were monitored

along the whole deposit thickness.

X-ray photoelectron and X-ray excited-Auger electron

spectroscopies (XPS and XE-AES) analyses were performed on

a PerkineElmerF 5600ci spectrometer at pressures lower than

10�8 mbar on a sample probing area of 1 mm � 1 mm, using

a MgKa excitation source (1253.6 eV). Binding Energy (BE)

values were corrected for charging by assigning to the C1s line

of adventitious carbon a value of 284.8 eV. Peak fitting was

performed by a least-squares procedure, using Gaus-

sianeLorentzian peak shapes. Zn and Ag Auger parameters

were calculated as previously reported [30,31,35].

Glancing incidence X-ray diffraction (GIXRD) measure-

ments were performed by a Bruker D8 Advance diffractometer

equipped with a Gobel mirror, using a CuKa source (40 kV,

40 mA), at a fixed incidence angle of 0.2�.

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 6 ( 2 0 1 1 ) 1 5 5 2 7e1 5 5 3 7 15529

Conventional transmission electron microscopy (TEM) and

high resolution (HRTEM) analyses were performed on cross-

section samples using a Tecnai G2 30 UT microscope with

a 0.17 nm point resolution, operated at 300 kV. Cross-section

specimens for TEM were prepared by mechanically grinding

down to the thickness of approximately 20 mm, followed by

Arþ ion beam milling using a Balzers RES 101 GVN apparatus.

High angle annular dark field-scanning transmission electron

microscopy (HAADF-STEM) and spatially resolved electron

energy loss spectroscopy (STEM-EELS) were performed on

a FEI Titan 80e300 “cubed” microscope fitted with an aberra-

tion corrector for the probe forming lens and a GIF QUANTUM

spectrometer, operated at 300 kV. The used HAADF detector

inner semi-angle, the convergence semi-angle for the EELS

experiments and the collection semi-angle were 90, 21, and

90 mrad, respectively. After acquisition, the EELS data were

treated using principle component analysis (PCA) to minimise

the effect of random noise. Maps were generated by inte-

grating the EELS signals over an appropriate energy window

after power-law background removal.

2.2. Hydrogen production and sensing

In order to prevent the occurrence of undesired alterations

upon functional testing, both hydrogen production and

sensing measurements were performed on ex-situ annealed

specimens. In fact, preliminary tests performed on as-

prepared samples resulted in the obtainment of irreproduc-

ible responses, that were related to the occurrence of

morphological alterations (system compaction) and compo-

sitional modifications (variations in the silver oxidation state).

UVeVis photocatalytic experiments were carried out in

a 250 mL pyrex discontinuous batch reactor maintained at

20 �C, filled with a 1:1 water-methanol mixture [2e4,8]. Ar gas

(15 mL � min�1) was used to collect and transfer gaseous

products to the analysis system (Fig. S1a, Supplementary

Content). A 125 W medium pressure mercury lamp (Helios

Italquartz) with pyrex walls was used for UVeVis excitation

[photon flux fi ¼ 29 mW � cm�2 (l ¼ 360 nm) and

100 mW � cm�2 (l ¼ 400O1050 nm), determined by a Del-

taOHM radiometer HD2302.0]. Experiments under simulated

solar radiation were carried out in a stainless steel photo-

reactor, using the same temperature and reaction mixture as

before (Fig. S1b, Supplementary Content). Illumination was

performed using a Solar Simulator (LOT-Oriel) equipped with

a 150 W Xe lamp, suitably filtered to reduce the fraction of UV

photons. The light intensity for simulated solar light experi-

ments, estimated by radiometry, was 180 mW � cm�2. Pho-

tocatalytic tests were repeated three times on different

samples to check reproducibility. On-line detection of evolved

gaseswas performed by anAgilent 6890 Gas Chromatographer

(GC), equipped with a 10 way-two loop injection valve. A

molecular sieve 5A column connected to a Thermal Conduc-

tivity Detector (TCD) was used for H2 analysis.

Gas sensing measurements were carried out at atmo-

spheric pressure in a sealed chamber maintained at 20 �C,using the flow-through technique. The sensor responses

[34,36] were investigated at temperatures between 100 and

400 �C, after pre-heating the samples at each operating

temperature for 8 h inside the test chamber for thermal

stabilization. Details on the system configuration, as well as

on the response, response time and recovery time definitions,

are reported in the Supplementary Content (see also Figs. S2-

S3, Supplementary Content).

The maximum recorded deviations for Ag/ZnO system

responses were lower than 10%, indicating a good reproduc-

ibility. As an example, typical variations for two subsequent

response scans to hydrogen (1000e5000 ppm concentrations,

200 �C; see Fig. 8b) were approximately 5%, a value within the

response uncertainty (compare Supplementary Content).

3. Results and discussion

3.1. Synthesis and characterization

In this work, three different sputtering times, leading to

different silver amounts, were adopted for Ag deposition onto

ZnO matrices (Table 1). Cross-section FE-SEM observations

showed that the zinc oxide columnar morphology did not

undergo significant alteration upon silver deposition, apart

from a systematic reduction of the column length from 180 to

120 nm (compare Fig. S4b, Supplementary Content). This

effect could be ascribed to the occurrence of ablation

phenomena, competitive with deposition ones in the used

plasmas [37]. Plane-view images of as-prepared Ag/ZnO

nanosystems (Fig. 1) clearly evidenced the presence of Ag

particles localized on the top of the columns, with morpho-

logical features directly dependent on the adopted sputtering

time. In particular, the mean Ag NP size progressively

increased on going from sample 30_A to 150_A, with

a concomitant shape evolution from spherical to island-like

aggregates. Upon annealing (Fig. 1), the formation of large

spherical Ag aggregates took place, due to thermally-induced

agglomeration of the pristineNPs (see Table 1), a phenomenon

particularly evident for sputtering times � 90 min.

The system composition as a function of the processing

conditions was investigated by EDXS along the deposit cross-

sections (Fig. 2). For all the analyzed specimens, the intensity

of O and Zn X-ray signals were always overlapped throughout

the sample depth, in line with the uniform and homogeneous

formation of zinc(II) oxide. In as-prepared samples, silver

presence was observed not only on the top of ZnO columns,

but also in the sub-surface layers. Such a behavior was

observed also for sample 150_A, characterized by the highest

silver loading, though in this case the strong intensity of the

Ag signal in the proximity of the surface suggested a metal

segregation in this region. Upon annealing, two different

behaviors were observed for Ag redistribution, depending on

the silver content. For sample 30_T, the dominant phenom-

enonwas an in-depth silver diffusion into ZnO. For specimens

90_T and 150_T, characterized by a progressively higher silver

content, the amount of silver diffused into the inner system

layers underwent an appreciable increase. In addition, excess

Ag segregated in the outermost region, forming spherical

particles on the sample surface. Note that a similar behavior

has also been observed upon annealing of Ag-implanted ZnO

samples [38,39].

To obtain a deeper insight into the system nanostructure,

with particular regard to the spatial distribution of the Ag NPs

Fig. 1 e Plane-view FE-SEMmicrographs of as-prepared and annealed Ag/ZnO nanocomposites. The presence of Ag particles

for sample 30_T is highlighted by a backscattered electron image (same scale bar) in the inset.

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 6 ( 2 0 1 1 ) 1 5 5 2 7e1 5 5 3 715530

inside the ZnO matrix, TEM analyses were carried out.

Representative cross-section TEM images of specimens 30_A

and 30_T are shown in Fig. 3a and b, respectively. At variance

with powdered ZnO (zincite), c-oriented zinc oxide systems

Fig. 2 e Cross-section EDXS line-scans for as-prepared and anne

correspond to O, Zn, and Ag signals, respectively (For interpret

reader is referred to the web version of this article.)

with (001) preferential orientation were obtained, as often

observed for ZnO nanostructures [9,15,17,22,24,26]. In the

present case, this [001] growth, leading to columns perpen-

dicular to the substrate surface, was further enhanced by the

aled Ag/ZnO nanocomposites. The red, green and blue lines

ation of the references to colour in this figure legend, the

Fig. 3 e Cross-section bright field TEM images of nanocomposites 30_A (a) and 30_T (b). White circles mark the location of Ag

NPs. (c) HRTEM of the deposit/Si(100) interface region highlighted by the black rectangle in (b). Note that Ag NPs are present

even at the interface with the substrate, where an amorphous SiO2 layer is also detected.

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 6 ( 2 0 1 1 ) 1 5 5 2 7e1 5 5 3 7 15531

sheath electric field [30,40]. In line with FE-SEM analyses,

sample 30_A showed small silver particlesmainly dispersed in

the outermost ZnO region (Fig. 3a). Conversely, for specimen

30_T (Fig. 3b) only a few Ag particles were present at the

surface, in agreement with the Ag in-depth distribution evi-

denced by EDXS. Accordingly, HRTEM images of sample 30_T

revealed the presence of small Ag NPs diffused through the

grain boundaries (GB) of the ZnO deposit, down to the inter-

face with the Si substrate (Fig. 3c). Such results unambigu-

ously confirm the in-depth penetration of silver along the

whole ZnO deposit, favoured by the porous ZnO morphology

and the typical infiltration power of RF-sputtering, and further

promoted by thermal treatment in air.

Fig. 4 displays typical HRTEM (a-b) and HAADF-STEM (c)

images for specimen 90_T. Beside the large silver aggregates

evidenced also by FE-SEM and EDXS (compare Figs. 1 and 2),

even smaller Ag NPs, with typical diameters between 5 and

20 nm, could be observed. Among them, low-sized Ag NPs

were single crystal and defect free, while larger ones showed

the presence of {111}-type twins (marked by white arrows in

Fig. 4a). Fig. 4b displays a ZnO column grown along the c-axis

in contact with a silver NP. Spatially resolved EELS measure-

ments were also performed on the region highlighted by the

box in Fig. 4c. As revealed by Fig. 4deg, elemental maps

confirm the presence of large silver particles segregated in the

outermost ZnO region.

Representative GIXRD patterns of Ag/ZnO systems are

displayed in Fig. 5. As a general rule, despite no significant

variation of ZnO peak positions was observed with respect to

the bare oxide nanosystems (compare Fig. S4e, Supplemen-

tary Content), the system structural features were influenced

by both the annealing process and the Ag deposition. In fact,

the appearance of additional diffraction peaks related to the

presence of metallic silver (JCPDS card no. 04-0783) or silver(I)

oxide (JCPDS card no. 72-2108) took place, whose intensity

increased with Ag sputtering time (compare patterns for 30_A

and 90_A). On going from sample 30_A to 30_T, the peak at

2q ¼ 38.1� ascribed to silver-containing phases disappeared, in

line with the enhanced silver particle dispersion into the ZnO

matrix evidenced by EDXS and TEM. Conversely, an appre-

ciable intensity increase of silver-related signals was detected

upon passing from specimen 90_A to 90_T, due to the previ-

ously discussed Ag segregation in the outermost system

regions. Furthermore, the appearance of peaks attributed to

hexagonal Ag2O (JCPDS card no. 72-2108) highlighted the

occurrence of silver oxidation during thermal treatment. At

variance with the results reported by Ramesh et al. for

annealed Ag2O films [41], no evidence of the Ag2O cubic phase

was obtained.

The surface chemical composition as a function of silver

content was investigated by photoelectron spectroscopies.

Irrespective of Ag sputtering time and annealing process, Zn

Auger parameter was always close to 2010.0 eV, the value

expected for ZnO [42]. Surface quantitative data (Table 1) show

that Ag content underwent a remarkable decrease after

thermal treatment, due to the concurrence of two different

phenomena: i) for 30_T, silver partially diffused into the inner

ZnO layers (see above); ii) for 90_T and 150_T specimens, the

reorganization of smaller Ag NPs into larger ones led to

a higher ZnO surface exposed to XPS probing.

Surface XPS Ag3d and XE-AES AgMVV spectra of a repre-

sentative Ag/ZnO nanocomposite are displayed in Fig. 6. For

all the investigated specimens, the Ag3d5/2 photopeak was

centered at a mean BE ¼ 368.0 eV. Due to the very close BE

values for Ag(0) and Ag(I) species [8,28,31,43e45], silver Auger

parameters were evaluated to obtain a finger print of silver

oxidation state [43]. For as-prepared samples, the obtained

values [a1(Ag) ¼ 724.3 eV; a2(Ag) ¼ 718.8 eV] revealed the

presence of Ag(I) oxide as the main silver species [31,35,43]. In

a different way, the increase of these values to

a1(Ag) ¼ 726.2 eV and a2(Ag) ¼ 720.3 eV upon annealing sug-

gested the occurrence of an Ag(I) reduction to Ag(0) [31,43].

This evolution could be related both to the thermal instability

of Ag(I) oxide and to themorphological reorganization of silver

aggregates, as highlighted by FE-SEM and EDXS observations.

In fact, Ag NPs underwent a significant size increase upon

annealing. It is reasonable to suppose that the smaller Ag

particles in as-prepared samples had a higher contact area

Fig. 4 e Cross-section TEM observations of specimen 90_T: (aeb) HRTEM micrographs; (c) HAADF-STEM image and

corresponding elemental maps (deg) extracted from STEM-EELS experiments. Color codes: Ag [ red, O [ green, Zn [ blue

(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

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 6 ( 2 0 1 1 ) 1 5 5 2 7e1 5 5 3 715532

with both ZnOmatrices and the outer atmosphere, which was

responsible for a more marked silver oxidation [27,28].

Conversely, the NP size increase and the Ag segregation in the

outermost ZnO region upon thermal treatment could explain

the predominance of Ag(0) at the sample surface. These

results suggest that the Ag2O phase observed in the GIXRD

patterns of annealed samples (Fig. 5) should be mainly local-

ized in the inner ZnO layers.

At variance with the case of pure ZnO samples (Fig. S4d,

Supplementary Content), three bands contributed to the O1s

signal for Ag sputtering times higher than 30 min (Fig. 6). The

average BE of the three components were: (I) 532.0 eV, (II)

530.5 eV, and (III) 529.4 eV. Whereas peak (I) could be ascribed

to surface chemisorbed carbonate and hydroxyl groups [42],

the position of peak (II) was in agreement with the values re-

ported for lattice oxygen in ZnO [30,42]. Component (III) could

be attributed to the presence of silver(I) oxide [28,43,46], as its

intensity increased with silver content.

Interestingly, a linear increase of the surface Zn2p3/2 BEs

with Ag content took place for the as-prepared samples

(Fig. S5, Supplementary Content), highlighting the occurrence

of electronic interactions between silver and zinc oxide. Even

for annealed samples, surface Zn2p3/2 BEswere systematically

higher than 1021.4 eV, the value expected for ZnO [42]. The

observed phenomenon could be related both to charge

transfer events from ZnO to Ag/Ag2O NPs [15,17], and to the

surface oxidation of Ag NPs to p-type Ag2O [46], resulting in

Ag2OeZnO p-n junctions generating a depletion layer at the

interface. The obtainment of Ag2OeZnO and AgeZnO inter-

faces with tunable features appears very promising for

a possible improvement of the system functional properties in

photocatalytic H2 generation and gas sensing.

3.2. Photo-assisted hydrogen production

Fig. 7a displays the rate of H2 photo-production vs. UVeVis

irradiation time for Ag/ZnO nanocomposites obtained under

different processing conditions.Whereas pure ZnO specimens

did not show any significant activity, an appreciable H2

evolution wasmeasured for Ag/ZnO systems, highlighting the

key role of silver NPs in promoting the photo-reforming of

methanol/water solutions under UVeVis irradiation. Under

Inte

nsit

y (a

. u.)

5045403530

2 (°)

ZnO

(00

2)

Ag

(111

) / A

g2O

(01

1)

30_T

30_A

Inte

nsit

y (a

. u.)

5045403530

2 (°)

ZnO

(00

2)

Ag

(111

)

Ag

(200

)Ag

2O (

011)

Ag

2O

(00

2) /

ZnO

(11

1)

90_A

90_T

Fig. 5 e GIXRD patterns of representative Ag/ZnO

nanocomposites before and after thermal treatment.

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 6 ( 2 0 1 1 ) 1 5 5 2 7e1 5 5 3 7 15533

these conditions, the process is likely to proceed through the

direct photoexcitation of electrons from the valence to the

conduction band of ZnO, followed by their transfer to Ag NPs

[8]. Subsequently, Hþ reduction on the metal particles results

in H2 generation, whereas alcohol oxidation is promoted by

holes in ZnO valence band. Beside the Schottky junction

between ZnO and Ag NPs, preventing electron-hole recombi-

nation [47], a further promotion of the observed processmight

result from the formation of Ag2OeZnO p-n junctions. A

similar effect has recently been observed for CuO-SnO2

nanocomposites [5].

As a matter of fact, H2 production rate was not directly

proportional to silver content, and the best performanceswere

observed for specimen 30_T. This phenomenon could be

explained taking into account that, despite Ag presence is

essential for the photo-assisted H2 evolution, a sufficiently

high fraction of uncovered ZnO surface must be available in

order to enable an efficient radiation absorption. When Ag NP

density is still limited and their size is low, such as for sample

30_T (see above and Fig. 1), the good photocatalytic perfor-

mances can be attributed both to an intimate contact between

Ag and ZnO and to a large exposed ZnO surface. In addition,

the highly dispersed Ag NPs in sample 30_T (compare SEM and

TEM results, Figs. 1e3) may favourably act as electron

capturing centers to generate H2. In a different way, for spec-

imen 90_T, the increased coverage of the ZnOmatrix by larger

Ag particles (compare Table 1 and Figs. 1 and 2) and the less

intimate AgeZnO contact resulted in a lowered radiation

absorption, explaining thus the observed trend. Upon

increasing the silver deposition time to 150 min, a similar

behavior was observed (specimen 150_T, data not shown).

The above results are consistent with those obtained for

Au/TiO2 nanosystems, whose optimal photocatalytic perfor-

mances were obtained for intermediate metal loading due to

a trade-off between gold NPs reactivity and their covering of

titanium oxide surface [48].

A careful inspection of Fig. 7a showed that, for the 30_T

composite, a stable activity was obtained, with an average H2

production rate of 0.4 mmol�1 � h�1 � cm�2. Conversely,

specimen 90_T presented a decrease of H2 production after 1 h

of irradiation. Such an effect could be related to the reduction

of Ag2O upon illumination [49], a process which is expected to

be less pronounced for smaller silver NPs, explaining thus the

good stability of sample 30_T.

In this contest, it is important to highlight that the

measurement of the Ag/ZnO exposed surface is a critical

issue. In general, conventional N2 or Kr physisorption

methods fail, as the estimation of the mass of the active

component with respect to the supporting substrate cannot

be performed with accuracy. On the other hand, scratching

of the deposited material from the substrate induces

further uncertainties. As a consequence, we preferred to

relate the activity to the geometrical surface, which has

indeed a practical meaning in terms of technological

applications.

A remarkable photo-reforming activity was observed

under illumination with simulated solar light (Fig. 7b), an

important result in view of practical utilization. Very inter-

estingly, H2 production rate over sample 90_T was more than

one order of magnitude higher than that reported in Fig. 7a,

where a more energetic radiation was used. Furthermore, at

variance with data obtained under UVeVis illumination,

a net performance enhancement was obtained upon passing

from 30_T to 90_T. Once again, no further improvement was

observed by increasing the deposition time to 150_T (data not

shown). These phenomena can be interpreted basing on the

combination of different effects. An increase of the metal NP

content can lead a more efficient Vis light absorption due to

the Surface Plasmon Resonance (SPR) of silver aggregates

[8,31,50]. Indeed, Ag SPR band centered at l z 410 nm was

observed for particles with diameters in the 5e100 nm range,

with a red-shift upon increasing metal NP size [51]. In addi-

tion, the presence of Ag2O species can also positively affect

the system reactivity. This observation is consistent with

previous findings for TiO2-based systems. In fact, Lalitha

et al. showed that pure TiO2 and Ag2O, or even Ag/TiO2

nanosystems, did not present high performances in photo-

assisted H2 production. Conversely, Ag2O/TiO2 nano-

composites yielded an improved efficiency, due to both

Inte

nsit

y (a

. u.)

910 905 900 895 890BE (eV)

AgM

5VV

AgM

4VV

Inte

nsit

y (a

. u.)

376 372 368 364BE (eV)

Ag3

d 5/2

Ag3

d 3/2

Inte

nsit

y (a

. u.)

535 530 525BE (eV)

(I)

(II)

30_A O1s

Inte

nsit

y (a

. u.)

535 530 525BE (eV)

(I)

(II)

30_T O1s

Inte

nsit

y (a

. u.)

535 530 525BE (eV)

(I)(II)

90_A

(III)

O1s

Inte

nsit

y (a

. u.)

535 530 525BE (eV)

(I)(II)

90_T

(III)

O1s

Fig. 6 e Surface Ag3d and AgMVV signals for an as-

prepared Ag/ZnO nanocomposite. Surface O1s photopeaks

for various Ag/ZnO nanocomposites.

0 1 2 3 4 5 6 7 80.0

0.2

0.4

0.6

0.8

30_T

90_T

ZnO

H2

evol

utio

n(µ

mol

×h-1

×cm

-2)

Time (h)

0 4 8 1 2 1 6 2 0 2 4 280

1

2

3

4

5

30_T

90_T

Time (h)

H2

evol

utio

n(µ

mol

×h-1

×cm

-2)

a

b

Fig. 7 e H2 evolution rate per unit area as a function of

illumination time on Ag/ZnO nanocomposites from H2O/

CH3OH solutions under: (a) UVeVis light; (b) simulated

solar irradiation.

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 6 ( 2 0 1 1 ) 1 5 5 2 7e1 5 5 3 715534

a beneficial interaction between Ag(I) species and TiO2, and

an enhanced radiation absorption in the Vis range [52].

Therefore, the presence of large silver aggregates and the low

accessibility of ZnO due to its high degree of coverage justify

the poorer reactivity of the 150_T sample with respect to the

90_T sample.

3.3. Hydrogen sensing

Finally, hydrogen sensing properties were investigated on

alumina-supportedAg/ZnOnanocomposites. For such systems,

GIXRDanalyses (not shown) revealed thepresence ofhexagonal

ZnO, with peaks of silver-containing phases detectable only for

sputtering times above 90 min. At variance with Si(100)-

supported specimens, the system morphology was character-

izedbyasortofurchin-likestructure [34], producedbyauniform

conformal coverage of globular Al2O3 grains by ZnO nano-

particles (mean lateral size¼ 20nm). Inaddition, thepresenceof

silver NPs on the oxide matrix could be observed (see Fig. 8a).

Dynamic response measurements (Fig. 8b) revealed

a conductance increaseuponhydrogen exposure, typical for n-

type semiconductors interacting with a reducing gas. It is

worth emphasizing that, irrespective of the preparative

conditions, the conductance variations upon injection of the

target gas were proportional to the H2 concentration, without

displaying any appreciable saturation effect. This behaviour,

alongwith thegoodrecoveryof theconductancebaselinevalue

upon switching off the analyte pulse, indicated a reversible H2

interactionwith theAg/ZnOsystem, an importantprerequisite

for the sensor technological application.

The present system responses were considerably high

(370-960 as the H2 concentration increased from 1000 to

5000 ppm) despite the moderate working temperature

(200 �C), and exceeded by far previous literature data for

hydrogen sensing by Ag/ZnO nanosystems [26]. The fast

response and recovery times (z3 min) were also promising in

view of practical applications. These performances can be

related to the morphology of ZnO nanoaggregates, presenting

a high surface-to-volume ratio, together with an average size

close to the zinc oxide Debye length [34], enabling their total

depletion under the test conditions (compare Fig. 8a).

Furthermore, Ag nanoparticles can exert a catalytic promo-

tion of the chemical processes involved in the sensing

Fig. 8 e (a) Plane-view FE-SEM micrograph of an annealed

Ag/ZnO sample deposited on polycrystalline Al2O3. A silver

particle is highlighted by an arrow. (b) Representative

dynamic response for a Ag/ZnO nanocomposite to square

H2 concentration pulses (working temperature [ 200 �C).

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 6 ( 2 0 1 1 ) 1 5 5 2 7e1 5 5 3 7 15535

mechanism, and improve charge separation at the interface

with ZnO [26].

4. Conclusions

This work was focused on a two-step approach to Ag/ZnO

nanocomposites combining PE-CVD, for the growth of

columnar ZnO arrays, and RF-sputtering, for the dispersion of

Ag NPs in the oxide matrices. Various sputtering times (from

30 to 150 min) were used to tailor the overall silver amount.

Depending on the adopted experimental conditions, the

processes enabled a fine tuning of Ag/ZnO nanocomposite

physico-chemical properties. In fact, the use of RF-sputtering

technique allowed a good silver dispersion into ZnOmatrices,

and the simultaneous control of Ag NPs size by variations of

the sole sputtering time. In addition, silver distribution in the

ZnO systems could be further controlled by thermal treat-

ments which, in turn, governed the Ag(0)-Ag(I) inter-conver-

sion leading to an electronic interplay between the different

phases. Such an effect directly impacted the system activity in

H2 production by photo-reforming, that yielded very attractive

and stable responses even upon activation with simulated

solar light. Due to AgeZnO interactions, very good perfor-

mances were also obtained in the gas-phase detection of H2.

Overall, the results provided herein pave the way to engi-

neering of the present nanosystems as multi-functional

platforms for both hydrogen generation and sensing.

Acknowledgements

These research activities have received funding from the Euro-

pean Community’s 6th Framework Program (FP7/2002-2006,

grant agreement n� Integrated Infrastructure Initiative, 026019,

ESTEEM) and 7th Framework Program (FP7/2007-2013, grant

agreement n� ENHANCE-238409). COFIN-PRIN 2008, Padova

University PRAT 2008/2010, Trieste University, Fondazione

CRTrieste, Regione Lombardia-INSTM PICASSO and Regione

Friuli Venezia Giulia IDROBIM projects are also acknowledged

for financial support. S.T. gratefully acknowledges the Fund for

Scientific Research Flanders (FWO). The Titanmicroscope used

in this studywas partially financed by theHercules Foundation.

Dr. Ke Xu and Manish Banerjee (Ruhr-University Bochum) are

gratefully acknowledged for technical assistance in the Zn

precursor synthesis.

Appendix. Supplementary material

Supplementary material associated with this article can be

found, in theonlineversion, atdoi:10.1016/j.ijhydene.2011.09.045

r e f e r e n c e s

[1] Muradov NZ, Veziro�glu TN. "Green" path from fossil-based tohydrogen economy: an overview of carbon-neutraltechnologies. Int J Hydrogen Energy 2008;33:6804e39.

[2] Alenzi N, Liao W-S, Cremer PS, Sanchez-Torres V, Wood TK,Ehlig-Economides C, et al. Photoelectrochemical hydrogenproduction from water/methanol decomposition using Ag/TiO2 nanocomposite thin films. Int J Hydrogen Energy 2010;35:11768e75.

[3] Miwa T, Kaneco S, Katsumata H, Suzuki T, Ohta K, ChandVerma S, et al. Photocatalytic hydrogen production fromaqueous methanol solution with CuO/Al2O3/TiO2

nanocomposite. Int J Hydrogen Energy 2010;35:6554e60.[4] Wu G, Chen T, Su W, Zhou G, Zong X, Lei Z, et al. H2

production with ultra-low CO selectivity via photocatalyticreforming of methanol on Au/TiO2 catalyst. Int J HydrogenEnergy 2008;33:1243e51.

[5] Zheng X-J, Wei Y-J, Wei L-F, Xie B, Wei M-B. PhotocatalyticH2 production from acetic acid solution over CuO/SnO2

nanocomposites under UV irradiation. Int J Hydrogen Energy2010;35:11709e18.

[6] Sahaym U, Norton M. Advances in the application ofnanotechnology in enabling a ‘hydrogen economy’. J MaterSci 2008;43:5395e429.

[7] Bak T, Nowotny J, Rekas M, Sorrell CC. Photo-electrochemicalhydrogen generation from water using solar energy.Materials-related aspects. Int J Hydrogen Energy 2002;27:991e1022.

[8] Liu J, Sun Y, Li Z, Li S, Zhao J. Photocatalytic hydrogenproduction from water/methanol solutions over highlyordered AgeSrTiO3 nanotube arrays. Int J Hydrogen Energy2011;36:5811e6.

[9] Wang ZL. Nanostructures of zinc oxide. Mater Today 2004;7:26e33.

[10] Ozgur U, Hofstetter D, Morkoc H. ZnO devices andapplications: a review of current status and future prospects.Proc IEEE 2010;98:1255e68.

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 6 ( 2 0 1 1 ) 1 5 5 2 7e1 5 5 3 715536

[11] Fu Q, Wagner T. Interaction of nanostructured metaloverlayers with oxide surfaces. Surf Sci Rep 2007;62:431e98.

[12] Zeng H, Cai W, Liu P, Xu X, Zhou H, Klingshirn C, et al. ZnO-based hollow nanoparticles by selective etching: eliminationand reconstruction of metal�semiconductor interface,improvement of blue emission and photocatalysis. ACSNano 2008;2:1661e70.

[13] Barreca D, Gasparotto A, Tondello E. Metal/oxide interfacesin inorganic nanosystems: what’s going on and what’s next?J Mater Chem 2011;21:1648e54.

[14] Height MJ, Pratsinis SE, Mekasuwandumrong O,Praserthdam P. AgeZnO catalysts for UV-photodegradationof methylene blue. Appl Catal B 2006;63:305e12.

[15] Lu W, Liu G, Gao S, Xing S, Wang J. Tyrosine-assistedpreparation of Ag/ZnO nanocomposites with enhancedphotocatalytic performance and synergistic antibacterialactivities. Nanotechnology 2008;19(445711):1e10.

[16] Tan T, Li Y, Liu Y, Wang B, Song X, Li E, et al. Two-step,preparation of Ag/tetrapod-like ZnO with photocatalyticactivity by thermal evaporation and sputtering. Mater ChemPhys 2008;111:305e8.

[17] Lin D, Wu H, Zhang R, Pan W. Enhanced photocatalysis ofelectrospun AgeZnO heterostructured nanofibers. ChemMater 2009;21:3479e84.

[18] Irimpan L, Nampoori VPN, Radhakrishnan P. Spectral andnonlinear optical characteristics of nanocomposites ofZnOeAg. Chem Phys Lett 2008;455:265e9.

[19] Zhang Y, Mu J. One-pot synthesis, photoluminescence, andphotocatalysis of Ag/ZnO composites. J Colloid Interface Sci2007;309:478e84.

[20] Roy RK, Bandyopadhyaya S, Pal AK. Surface plasmonresonance in nanocrystalline silver in a ZnO matrix. Eur PhysJ B 2004;39:491e8.

[21] Prokes SM, Glembocki OJ, Rendell RW, Ancona MG.Enhanced plasmon coupling in crossed dielectric/metalnanowire composite geometries and applications to surface-enhanced Raman spectroscopy. Appl Phys Lett 2007;90(093105):1e3.

[22] Deng S, Fan HM, Zhang X, Loh KP, Cheng C-L, Sow CH, et al.An effective surface-enhanced Raman scattering templatebased on a Ag nanocluster-ZnO nanowire array.Nanotechnology 2009;20(175705):1e7.

[23] Mo L, Zheng X, Yeh C-T. Selective production of hydrogenfrom partial oxidation of methanol over silver catalysts atlow temperatures. Chem Comm 2004:1426e7.

[24] Perez-Hernandez R, Gutierrez-Martınez A, Mayoral A,Deepak FL, Fernandez-Garcıa ME, Mondragon-Galicia G, et al.Hydrogen production by steam reforming of methanol overa Ag/ZnO one dimensional catalyst. Adv Mater Res 2010;132:205e19.

[25] Yao W, Huang C, Muradov N, T-Raissi A. A novel PdeCr2O3/CdS photocatalyst for solar hydrogen production usinga regenerable sacrificial donor. Int J Hydrogen Energy 2011;36:4710e5.

[26] Xiang Q, Meng G, Zhang Y, Xu J, Xu P, Pan Q, et al. Agnanoparticle embedded-ZnO nanorods synthesized viaa photochemical method and its gas-sensing properties.Sens Actuators B 2010;143:635e40.

[27] Qi H, Alexson D, Glembocki O, Prokes SM. The effect of sizeand size distribution on the oxidation kinetics andplasmonics of nanoscale Ag particles. Nanotechnology 2010;21(215706):1e5.

[28] Kibis LS, Stadnichenko AI, Pajetnov EM, Koscheev SV,Zaykovskii VI, Boronin AI. The investigation of oxidizedsilver nanoparticles prepared by thermal evaporation andradio-frequency sputtering of metallic silver under oxygen.Appl Surf Sci 2010;257:404e13.

[29] Arabatzis IM, Stergiopoulos T, Bernard MC, Labou D,Neophytides SG, Falaras P. Silver-modified titanium dioxidethin films for efficient photodegradation of methyl orange.Appl Catal B 2003;42:187e201.

[30] Bekermann D, Gasparotto A, Barreca D, Bovo L, Devi A,Fischer RA, et al. Highly oriented ZnO nanorod arrays bya novel plasma chemical vapor deposition process. CrystGrowth Des 2010;10:2011e8.

[31] Barreca D, Gasparotto A, Maragno C, Tondello E, Gialanella S.Silica-supported silver nanoparticles: tailoring of structure-property relationships. J Appl Phys 2005;97(054311):1e8.

[32] Zheng J, Yang R, Xie L, Qu J, Liu Y, Li X. Plasma-assistedapproaches in inorganic nanostructure fabrication. AdvMater 2010;22:1451e73.

[33] Bekermann D, Rogalla D, Becker H-W, Winter M, Fischer RA,Devi A. Volatile, monomeric, and fluorine-free precursors forthe metal organic chemical vapor deposition of zinc oxide.Eur J Inorg Chem; 2010:1366e72.

[34] Barreca D, Bekermann D, Comini E, Devi A, Fischer RA,Gasparotto A, et al. 1D ZnO nano-assemblies by plasma-CVDas chemical sensors for flammable and toxic gases. SensActuators B 2010;149:1e7.

[35] Armelao L, Barreca D, Bottaro G, Gasparotto A, Maccato C,Tondello E, et al. Rational Design of Ag/TiO2 nanosystems bya combined RF-Sputtering/Sol-Gel approach.ChemPhysChem 2009;10:3249e59.

[36] Barreca D, Gasparotto A, Maccato C, Maragno C, Tondello E,Comini E, et al. Columnar CeO2 nanostructures for sensorapplication. Nanotechnology 2007;18(125502):1e6.

[37] Barreca D, Carraro G, Gasparotto A, Maccato C, Lebedev OI,Parfenova A, et al. Tailored vapor-phase growth ofCuxOeTiO2 (x ¼ 1, 2) nanomaterials Decorated with Auparticles. Langmuir 2011;27:6409e17.

[38] Rita E, Wahl U, Lopes AML, Araujo JP, Correia JG, Alves E,et al. Lattice site and stability of implanted Ag in ZnO.Physica B 2003;340-342:240e4.

[39] Sakaguchi I, Watanabe K, Ohgaki T, Nakagawa T,Hishita S, Adachi Y, et al. Ion implantation and diffusionbehavior of silver in zinc oxide. J Ceram Soc Jpn 2010;118:217e9.

[40] Romero-Gomez P, Toudert J, Sanchez-Valencia JR,Borras A, Barranco A, Gonzalez-Elipe AR. Tunablenanostructure and photoluminescence of columnar ZnOfilms grown by plasma deposition. J Phys Chem C 2010;114:20932e40.

[41] Ramesh MNV, Sundarayya Y, Sunandana CS. Reactivelyradio frequency sputtered silver oxide thin films: phaseevolution and phase stability. Mod Phys Lett B 2007;21:1933e44.

[42] Barreca D, Gasparotto A, Maccato C, Maragno C, Tondello E.ZnO nanoplatelets obtained by chemical vapor deposition,studied by XPS. Surf Sci Spectra 2007;14:19e26.

[43] Waterhouse GIN, Bowmaker GA, Metson JB. Oxidation ofa polycrystalline silver foil by reaction with ozone. Appl SurfSci 2001;183:191e204.

[44] Michaelides A, Reuter K, Scheffler M. When seeing is notbelieving: oxygen on Ag(111), a simple adsorption system? JVac Sci Technol A 2005;23:1487e97.

[45] Sen S, Mahanty S, Roy S, Heintz O, Bourgeois S, Chaumont D.Investigation on sol-gel synthesized Ag-doped TiO2 cermetthin films. Thin Solid Films 2005;474:245e9.

[46] Gao X-Y, Feng H-L, Ma J-M, Zhang Z-Y, Lu J-X, Chen Y-S, et al.Analysis of the dielectric constants of the Ag2O film byspectroscopic ellipsometry and single-oscillator model.Physica B 2010;405:1922e6.

[47] Primo A, Corma A, Garcia H. Titania supported goldnanoparticles as photocatalyst. Phys Chem Chem Phys 2011;13:886e910.

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 6 ( 2 0 1 1 ) 1 5 5 2 7e1 5 5 3 7 15537

[48] Armelao L, Barreca D, Bottaro G, Gasparotto A, Maccato C,Maragno C, et al. Photocatalytic and antibacterial activity ofTiO2 and Au/TiO2 nanosystems. Nanotechnology 2007;18(375709):1e7.

[49] Kudo A, Miseki Y. Heterogeneous photocatalyst materials forwater splitting. Chem Soc Rev 2009;38:253e78.

[50] Chuang H-Y, Chen D-H. Fabrication andphotoelectrochemical study of Ag@TiO2 nanoparticle thinfilm electrode. Int J Hydrogen Energy 2011;36:9487e95.

[51] Zielinska A, Skwarek E, Zaleska A, Gazda M, Hupka J.Preparation of silver nanoparticles with controlled particlesize. Procedia Chem 2009;1:1560e6.

[52] Lalitha K, Reddy JK, Sharma MVP, Kumari VD,Subrahmanyam M. Continuous hydrogen productionactivity over finely dispersed Ag2O/TiO2 catalysts frommethanol:water mixtures under solar irradiation:a structure-activity correlation. Int J Hydrogen Energy 2010;35:3991e4001.