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 9 ( 2 0 1 4 ) 7 2 4 7e7 2 6 1

Available online at w

ScienceDirect

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

Review

Photocatalytic membrane reactors for hydrogenproduction from water

Raffaele Molinari a,*, Tiziana Marino b, Pietro Argurio a

aDepartment of Environmental and Chemical Engineering, University of Calabria, Via Pietro Bucci,

cubo 44/A, I-87036 Rende (CS), ItalybNational Research Council e Institute on Membrane Technology (ITMeCNR), c/o University of Calabria,

Via Pietro Bucci, cubo 17C, I-87036 Rende (CS), Italy

a r t i c l e i n f o

Article history:

Received 7 November 2013

Received in revised form

21 February 2014

Accepted 28 February 2014

Available online 30 March 2014

Keywords:

Photocatalytic membrane reactors

Hydrogen production from water

Overall water splitting

* Corresponding author.E-mail addresses: raffaele.molinari@unic

http://dx.doi.org/10.1016/j.ijhydene.2014.02.10360-3199/Copyright ª 2014, Hydrogen Ener

a b s t r a c t

Hydrogen is considered today a promising environmental friendly energy carrier for the

next future, since it produces no air pollutants or greenhouse gases when it burns in air,

and it possesses high energy capacity. In the last decades great attention has been devoted

to hydrogen production from water splitting by photocatalysis. This technology appears

very attractive thanks to the possibility to work under mild conditions producing no

harmful by-products with the possibility to use renewable solar energy. Besides, it can be

combined with the technology of membrane separations making the so-called photo-

catalytic membrane reactors (PMRs) where the chemical reaction, the recovery of the

photocatalyst and the separation of products and/or intermediates simultaneously occur.

In this work the basic principles of photocatalytic hydrogen generation fromwater splitting

are reported, giving particular attention on the use of modified photocatalysts able to work

under visible light irradiation. Several devices to achieve the photocatalytic hydrogen

generation are presented focusing on the possibility to obtain pure hydrogen employing

membrane systems and visible light irradiation. Although many efforts are still necessary

to improve the performance of the process, membrane photoreactors seem to be promising

for hydrogen production by overall water splitting in a cost-effective and environmentally

sustainable way.

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

reserved.

Introduction

Hydrogen is considered an ideal fuel for the next future. It

does not produce any contaminant when it burns in air, it is a

clean and a non-polluting energy carrier. It is the most

al.it, r.molinari@unical.it

74gy Publications, LLC. Publ

common element on earth, mostly present in water, biomass

and hydrocarbons. Nowadays, the most widely used indus-

trial process for hydrogen production is the Steam Methane

Reforming (SMR), which is also the most economical process

[1]. This process involves many different catalytic steps, as

long as natural gas (or methane) and hydrocarbon fuels that

(R. Molinari).

ished 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 9 ( 2 0 1 4 ) 7 2 4 7e7 2 6 17248

remains at a low or moderate price. However, industrial

hydrogen production results in equally large carbon dioxide

emissions. Moreover, the emission of greenhouse gases also

needs to be reduced to solve the problem of the global

warming.

Other fuel processing technologies for hydrogen produc-

tion are partial oxidation and autothermal reforming. Partial

oxidation converts hydrocarbons to hydrogen by partially

combusting the hydrocarbon with oxygen. The process re-

quires high temperatures with some soot formation and the

hydrogen/carbon oxide ratio from 1:1 to 2:1. Autothermal

reforming is typically conducted at a lower pressure than

partial oxidation reforming and has a low methane slip to

atmosphere. However, these processes require an expensive

and complex oxygen separation unit in order to feed pure

oxygen to the reactor.

Fuel fossil burning, which contributes to the greenhouse

gas pool, and also the eventual depletion of the world’s fossil

fuel reserves need for a development of cheap and very effi-

cient new technologies. One strategy could be to apply steam

reforming methods to alternative renewable materials [2e4]

such as that ones from crops. Not only do these biomass-

conversion schemes turn waste into a valuable product, but

in addition, any carbon dioxide in the processes could be

soaked up by planting new crops to provide the required

biomass. A biomass strategy could represent a useful alter-

native to the current fossil fuel methods.

Solar and wind that are the two major sources of renew-

able energy, also represent a promising way for sustainable

hydrogen production, although their cost still remains high.

In the last years there has been a great interest of research in

water splitting to make hydrogen using thermochemical,

photobiological and photocatalytic systems. In thermo-

chemical processes, heat from sunlight at around 2000 �Ccould be collected and used to carry out the water photode-

composition in presence of a semiconductor. The main

drawback of this approach is the careful control of materials

able to resist to heat, in addition to the high cost required to

concentrate the solar light [5]. Photobiological water splitting

is based on the use of oxygenic/anoxygenic photosynthetic

microorganisms in anaerobic conditions. This technology

could be an efficient method to generate hydrogen also from

waste water but it still presents the problem of the carbon

dioxide formation in the products when anoxygenic bacteria

are used to promote the reaction. Hydrogen production by

oxygenic bacteria currently offer more advantages because it

leads to the decomposition of water into hydrogen and oxy-

gen, but it also has many difficulties especially for the scale-

up and the low hydrogen yield. Water electrolysis could

become a useful alternative for producing clean hydrogen. It

is essentially the conversion of electrical energy to chemical

energy in the hydrogen form, with oxygen as useful by-

product [6]. The most common electrolysis technology is

alkaline based, but the research is developing also in proton

exchange membrane processes and solid oxide electrolysis

cells units. Currently, electrolysis is more expensive than

using large-scale fuel processing techniques to produce

hydrogen. It can become more competitive as the cost con-

tinues to decrease with the technology advancement. Alter-

natively, photocatalytic water splitting using

semiconductors could offer a promising way for low cost and

environmentally friendly hydrogen generation from solar

energy [7e19].

In the last decade photocatalysis has become more and

more attractive for the research in water splitting to produce

hydrogen and oxygen using clean and renewable sources, for

organic synthesis and for purification of water and air with

benefit for environment and for the industry regarding the

development of technologies. Compared with traditional

advanced oxidation processes the technology of photo-

catalysis is known to have some advantages, such as easy

setup and operation at ambient temperatures, no need for

post-processes, low consumption of energy and consequently

low costs.

Increasing attention has being taken to recently developed

photocatalytic membrane reactors (PMRs), devices which

combine a photocatalytic process with a membrane separa-

tion to obtain chemical transformations [20]. PMRs improves

the potentialities of classical photoreactors (PRs) and those of

membrane processes [21e23] with a synergy of both tech-

nologies thus minimizing environmental and economical

impacts [24]. The membrane permits continuous operation in

systems in which the recovery of the photocatalyst, the re-

action and the products separation simultaneously occur.

Higher energy efficiency, modularity and easy scale-up are

some other advantages of PMRs with respect to convectional

PRs.

In this paper the basic principles of photocatalytic

hydrogen generation are presented, focusing on the photo-

catalysts able to work under visible light irradiation. Themain

techniques to modify semiconductors and their applications

in the photodecomposition of water are reviewed. The

development of membrane reactors able to work under

visible light irradiation for the pure hydrogen generation is

reported.

Basics principles of photocatalytic hydrogengeneration

Heterogeneous photocatalysis was defined by Palmisano and

Sclafani [25] in 1997 as “a catalytic process during which one

or more reaction steps occur by means of electronehole pairs

photogenerated on the surface of semiconducting materials

illuminated by light of suitable energy”. As a direct conse-

quence of this definition, both catalyst and light are necessary

to induce a chemical process. In fact, upon irradiation excited

states of the photocatalyst are generated and initiate subse-

quent processes like reductioneoxidation reactions and mo-

lecular transformations. The basic mechanism of

heterogeneous photocatalysis have been investigated by

many research groups [26,27] and can be summarized by the

Fig. 1.

The irradiation of a semiconductor with light of energy

equal or higher than its band gap energy (Eg) gives rise to

promotion of electrons (e�) from the valence band to the

conduction band, leaving at the same time positive holes (hþ)in the valence band. The photogenerated electronehole pairs

can induce redox reactions with electron donor (D in Fig. 1)

and electron acceptor (A in Fig. 1) adsorbed on the

Fig. 1 e Simplified reaction scheme of photocatalysis.

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 2 4 7e7 2 6 1 7249

semiconductor surface or located within the electrical double

layer surrounding the charged particles. The oxidation-

reduction process competes with a possible electrons-holes

recombination (within a few nanoseconds), with a conse-

quent input energy dissipation as heat. A thermodynamic

condition necessary to the occurrence of such a process is a

more negative potential for the electrons photogenerated in

the conduction band comparedwith the reduction potential of

the electron acceptor and a more positive potential for the

holes of the valence band compared with the oxidation po-

tential of the donor.

The potentials of both conduction and valence bands are

pH dependent. The increase of pH in a electrolyte solution

causes a shift of conduction and valence band potential to-

wards more cathodic values (59 mV per pH unit) [28]. Among

all investigated semiconductors (generally metal oxides and

chalcogenides) TiO2 is by far the most largely used [29].

Since Fujishima and Honda reported photoelectrochemical

water splitting using a TiO2 electrode in 1972 [30], many

research groups have intensively studiedwater splitting using

photoelectrodes or photocatalysts [29]. Both photocatalytic

Fig. 2 e Photocatalytic hydrogen generation upon a

suitable semiconductor.

and photoelectrochemical water splitting are similar to the

photosynthesis process of green plants, so are both inter-

esting and promising topics. However, heterogeneous photo-

catalysis that employees photocatalyst powders has several

advantages over photoelectrochemical cells such as greater

simplicity and lower processing cost. In both systems the

electronic structure of a semiconductor plays a key role for an

efficient hydrogen generation [29]. The photocatalytic

hydrogen generation using a useful semiconductor is repre-

sented in Fig. 2.

For hydrogen generation, the conduction band level of the

semiconductor should be more negative than hydrogen pro-

duction level, while the valence band should be more positive

than water oxidation level to obtain an efficient oxygen gen-

eration from photocatalysis. In the overall water splitting

process, oxygen and hydrogen evolution can be modeled as

two chemical half reactions. A simplified equation set that

describes these half reactions in addition to the net conver-

sion process can be written:

2g / 2e�CB þ 2hþVB Photon induced e�/hþ generation (1)

H2O þ 2hþ / 2Hþ þ 1/2O2 Water oxidation half reaction (2)

2Hþ þ 2e� / H2 (gas) Proton reduction half reaction (3)

H2O þ 2g / H2 (gas) þ 1/2 O2 Net overall water splitting (4)

DG� ¼ þ 237.18 kJ mol�1 Standard Gibbs free energy (5)

V�rev ¼ DG�/nF ¼ 1.23 V Standard reversible potential (6)

where g is photon energy, e� is an electron, hþ is a hole, DG�

is the standard Gibbs free energy, V�rev is the standard

reversible potential, n( ¼ 2) is the number of electron

exchanged and F is the Faraday constant. It is clear from the

equation set that photocatalytic water splitting is a delicate

balancing process, where photoproduced electronehole pairs

can drive the photocatalytic half reactions. In this system,

light energy is converted into chemical energy. Thus, the

minimum energy required to drive the reaction referred to

that of two moles of impinging photons is equal to 1.23 V.

Presently, hydrogen generation fromwater splitting has still a

low efficiency, mainly due to the recombination of the pho-

togenerated electronehole pairs, the inability to use visible

light using most of known photocatalysts and in many cases

to the possible reverse reaction that involves the rapid

hydrogen and oxygen recombination. Since solar energy

incident on the Earth’s surface is composed for a large part of

the visible region the use of visible light to promote hydrogen

generation from water splitting represents an extremely

interesting target. Even if all UV light and visible light up to

800 nm is used a conversion efficiency of w30% could be ob-

tained [16].

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 2 4 7e7 2 6 17250

Photocatalysts for hydrogen generation

Heterogeneous photocatalysts are solids able to promote re-

actions in the presence of light without being consumed in the

overall reaction. Catalysts used in photocatalytic reactions are

invariably semiconductor materials. Chalcogenide include

several metal oxides such as TiO2, CeO2, ZnO, WO3, Fe2O3 or

even sulfides like CdS, ZnS, MoS2 [31e35]. Having high pho-

tocatalytic activity and chemical stability, non-toxicity and

low cost, titanium dioxide has become the most widely used

photocatalyst. It has been widely used as white pigment in

paints, plastic, paper, cosmetics and foodstuffs. Three crys-

talline forms exist in nature: anatase (tetragonal), rutile

(tetragonal) and brookite (orthorhombic). Among them,

anatase has been found to be the most photocatalytically

reactive form [36] even if the more stable form is rutile,

whereas anatase and brookite are metastable and are readily

transformed to rutile when heated [37]. TiO2 has a band gap

energy of w3.2 eV, hence its photoactivation requires radia-

tion with light of wavelength less than or equal to w380 nm,

with a maximum at approximately 340 nm. One of the most

used commercial TiO2 materials for photocatalytic oxidation

applications is TiO2 Degussa P25 (now Evonik). It is composed

by amixture of 80% anatase and 20% rutile, with a surface area

of 50 m2 g�1 and an average particle size of 21 nm (manufac-

turer data). Sensitized photoreactions are activated by ab-

sorption of a photon with sufficient energy, i.e., equal or

higher than the band gap energy of the catalyst. Band gap

energies of the most common semiconductors are shown in

Table 1.

Nevertheless, some of the photocatalysts do not show a

long-term activity during the irradiation process. Binarymetal

sulfides such as CdS, CdSe or PbS rapidly undergo photoanodic

corrosion so they are insufficiently stable for catalysis in

aqueous media [38,39]. To overcame this drawback the addi-

tion of sulfide and sulfite to the contacting solution is sug-

gested in several works [40]. These materials are also known

Table 1 e Band gap energies of semiconductors used inphotocatalytic processes.

Photocatalyst Band gap energy (eV)

Si 1.1

WSe2 1.2

WO3 2.8

a-Fe2O3 2.2

V2O5 2.7

SiC 3.0

BaTiO3 3.3

CdO 2.1

CdS 2.4

CdSe 1.7

Fe2O3 3.1

TiO2 rutile 3.0

TiO2 anatase 3.2

SrTiO3 3.4

SnO2 3.5

GaAs 1.4

SrTiO3 3.4

ZnS 3.7

ZnO 3.2

to be toxic. Hematite (a-Fe2O3), for example, is absorptive in

the visible region, but shows much lower photocatalytic ac-

tivity than TiO2 or ZnO, probably because of corrosion or the

formation of short-lived metal-to-ligand or ligand-to-metal

charge transfer states [41]. Although ZnO and TiO2 (anatase)

have similar band gap energies (3.2 eV) and higher reactivity

than TiO2 (rutile) [42], zinc oxide lead to Zn(OH)2 formation in

illuminated aqueous solutions, favoring the catalyst deacti-

vation [43].

Wu [44] also observed higher photocatalytic activity for

TiO2 compared to SnO2. WO3 is a powerful semiconductor for

many types of photocatalytic processes [45e47] having a

remarkable photostability in acidicmedia. It has a band gap of

w2.8 eV, that is w0.4 eV narrower than that of TiO2, so it can

absorb more visible light from sunlight [48]. As one kind of

important rare earth materials CeO2 in the last years has

attracted great interest for its potential photocatalytic appli-

cations [49,50]. Furthermore, nanocrystalline cerium dioxide

has some singular properties like as transition from boundary

diffusion to lattice diffusion [51] and lattice expansion [52].

In the next section, a detailed description about TiO2 and

CeO2 photocatalysts is presented. Ceria nanoparticles show a

strong absorption below 400 nmwith a maximum absorbance

peak around 305 nm. The band gap energy varies in the range

3.03e3.7 eV depending on the preparation method [52].

Modified photocatalysts

The efficiency of the classical photocatalysts is often reduced

by some drawbacks, such as the very quickly recombination of

the photogenerated electronehole pairs (within 10e100 ns)

which releases thermal energy or unproductive photons, the

possibility of backward or secondary reactions with the for-

mation of undesirable by-products and a low absorption in the

visible region which determines their inability to use solar

light. In recent years, therefore, the development of photo-

catalysts able to overcome these problems represents one of

the main topic in the photocatalytic research. The major

practices involve catalystmodification by noblemetal loading,

ion doping into the semiconductor lattice, dye photosensiti-

zation and mixing with other semiconductors.

Noble metals like Pt, Pd, Au, Rh, Ni, Cu and Ag have been

reported to be very effective to enhance catalytic activity of

several semiconductors and the visible light absorption

[53e55]. Noble metals doped or deposited on a photocatalyst

are expected to show various effects on its photocatalytic

activity by different mechanisms. They are able to enhance

the electronehole separation by acting as electron traps, to

extend the light absorption into the visible range and to

modify the surface properties of photocatalyst [56]. As one of

the noble metals, Pt has been widely used as the cocatalyst in

photocatalytic water splitting over many different kinds of

semiconductors: oxides [57], (oxy)sulfides [58] and (oxy)ni-

trides [59]. All have been shown to greatly enhance the pho-

tocatalytic activity for hydrogen evolution. Wu et al. [60]

investigated hydrogen production with low CO selectivity

from the photocatalytic reforming of glucose in water on

metal/TiO2 catalysts (metal Pt, Rh, Ru, Ir, Au, Ni, and Cu). The

loadedmetals, in particular Rh, were found to greatly enhance

the rate of hydrogen production. This was attributed to the

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 2 4 7e7 2 6 1 7251

fact that the Schottky barrier formed at the metal-TiO2 inter-

face could serve as an efficient electron trap, preventing

photogenerated electronehole recombination. The effect of

the nature of the metal cocatalyst was interpreted in terms of

different electronic interactions between the metal nano-

particles and the TiO2 surface. It was also reported that the

smaller the Schottky barrier height at the metal/semi-

conductor junction, the greater was the electron flow from

semiconductor to metal, thus leading to higher photocatalytic

activity for hydrogen production [61]. When cocatalysts con-

sisting of Au/Pd bimetallic nanoparticles with core/shell

structures were loaded onto the TiO2 surface, Pd shell allowed

selective hydrogen permeability and contributed to reduce the

protons with photogenerated electrons. As result the photo-

catalytic production of hydrogen from aqueous ethanol solu-

tions was greatly enhanced [62].

Among noble metals and in view of potential commercial

applications, gold has been intensively studied in material

science, especially in the context of emerging nanotech-

nology, as also confirmed from the large number of papers

reviewing the topic [63,64]. Methods of preparation of the

semiconductor-metal catalyst play an important role in the

photocatalytic activity of the overall system. This aspect was

analyzed in detail by Silva et al. [65], which studied the pho-

tocatalytic activity of several Au/TiO2 samples on the

hydrogen formation from water, under different light irradi-

ation sources and sacrificial agents in the aqueous medium.

The size of the gold nanoparticles is key in obtaining active

catalysts for water splitting [65]. Most gold precursors, used in

preparing supported gold catalysts, are salt complexes where

the oxidation state of gold is typically þ3 since the latter is

more stable than the þ1 oxidation state. Moreover, the size of

the gold nanoparticles highly depend on the gold ion con-

centration, the pH and the temperature of the solution used.

Small gold nanoparticles with sizes of less than 10 nm can be

obtained by depositing gold onto a solid support.

Different methods for the synthesis of supported gold

nanoparticles have been reported in the literature to obtain

well-dispersed particles with nanometer sizes. The most

commonly used procedure to synthesize gold nanoparticles is

the deposition-precipitation (DP) method, suggested by Har-

uta [64]. This technique involves the deposition of gold hy-

droxide on the surface of the metal oxide support by raising

the pH of the gold chloride precursors. The pH of aqueous

HAuCl4 solution is adjusted at a fixed point in the range of

6e10, and is selected primarily based on the isoelectric points

of themetal oxide supports [64]. Careful control of the solution

pH enables selective deposition of Au(OH)3 only on the sur-

faces of the support metal oxides without precipitation in the

liquid phase. For the novel visible light photoactivity of Au/

TiO2, the authors determined that gold loading, particle size

and calcination temperature play a role in the photocatalytic

activity, the most active material (quantum efficiency for

hydrogen and oxygen generation 7.5% and 5.0%, respectively,

at 560 nm) being the catalyst containing 0.2 wt % gold with

1.87 nm average particle size and calcined at 200 �C.Rare earth metal ions and transitional metal ions doping

have found to be able for enhancing the photocatalytic activity

of several semiconductors [66,67]. CeO2 in the last years has

attracted great interest for its potential photocatalytic

applications [49]. A novel method to prepare ceria powder to

reduce the particle size of ceria (5 nm) by means of electro-

static binding of Ce4þ to alginate gel, subsequent supercritical

CO2 drying, and calcination was proposed by Primo et al. [68].

They showed that deposition of gold nanoparticles at low

loading increases the photocatalytic activity for visible light-

producing samples that exhibit higher photocatalytic activity

than the same material upon irradiation at its band gap.

Moreover, the ceria samples containing gold under visible

light irradiation outperform the photocatalytic activity ofWO3

under UV light irradiation. When irradiated with visible light

(l > 400 nm) 1.0 wt% gold-supported ceria nanoparticles

generate oxygen from water (10.5 mmol h�1) more efficiently

than the standard WO3 (1.7 mmol h�1) even under UV irradia-

tion (9.5 mmol h�1).

Over the past decades, there have been numerous reports

on the modification of wide band gap photocatalysts using

metal ion doping to make them visible light active. These

include doped TiO2 [69], SrTiO3 [70], ZnO [71] and doped ZnS

[72] among others. Themain objective of doping is to induce a

batho-chromic shift, i.e., a decrease of the band gap or intro-

duction of intra-band gap states, which results in more visible

light absorption. The dopant ions can work as hole and elec-

tron traps or they can mediate interfacial charge transfer.

Dopant ions may be adsorbed on the semiconductor catalyst,

incorporated into the interior of the semiconductor particle, or

they may form separate oxide phases. Once incorporated into

the interior of the photocatalyst, the dopant ions may occupy

either lattice (substitutional) or interstitial sites. Metal ions

such as Fe, Mo, Ru, Os, Re, V and Rh can enhance TiO2 pho-

toactivity, while others ions like Co and Al cause detrimental

effects. Kudo and Kato [73] investigated the effects of doping

lanthanide like La, Pr, Nd, Sm, Gd, Tb and Dy and alkaline

earth metal ions such as Ca, Sr, and Ba into NaTaO3 photo-

catalysts for efficient water splitting. Lanthanum was the

most effective dopant. The apparent quantum yield at 270 nm

amounted to 56% which is the highest quantum yield ever

reported for catalysts in pure water splitting. The positive ef-

fects on the photocatalytic properties were mainly due to the

decrease in the particle size and the ordered surface

nanostructure.

Dye sensitization of nanocrystalline semiconductors has

attracted considerable attention in the last years since

Graetzel first reported on the highly efficient ruthenium

complex sensitized nanocrystalline TiO2-based dye-sensi-

tized solar cells [74]. Upon illumination by visible light, the

excited dyes can inject electrons to the semiconductor con-

duction band to initiate the photocatalytic process. The

semiconductor surface acts as a quencher accepting an elec-

tron from the excited dye molecule. The electrons can in turn

be transferred to reduce an acceptor species adsorbed on the

surface. Dye-sensitized solar cells have attracted a great deal

of interest because of their relatively higher efficiency and low

cost compared with conventional inorganic photovoltaic de-

vices [75]. Dhanalakshmi et al. [76] found that, when TiO2 and

ZnO were sensitized with a new sensitizer ([Ru-

(dcbpy)2(dpq)]2þ), they displayed extremely stable and effi-

cient photocatalytic activity for hydrogen production under

visible light irradiation from water even in the absence of an

electron donor. Gratzel and co-workers [77] succeeded in

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 2 4 7e7 2 6 17252

decomposing water by visible light using Ru(bpy)32þ and its

amphiphilic derivatives as sensitizers. Pt/RuO2-loaded TiO2

particles proved particularly effective in those systems, acting

as photocatalysts for water splitting process. Nada et al. [78]

found that, when copper phthalocyanine was used as a

photosensitizer, it was effective for hydrogen production over

RuO2/TiO2 using UV/solar light irradiation. In fact, copper

phthalocyanine exhibited higher efficiency compared to other

sensitizers such as ruthenium bipyridyl.

Latorre-Sanchez et al. [79] studied the photocatalytic effi-

ciency of graphene oxide for hydrogen generation from water

using methanol as sacrificial agent in presence of dyes. This

studywas based on two graphene oxidematerials obtained by

chemical oxidation of graphite followed by ultra-sound soni-

cation. The twomaterials differ between themby about 10% in

their carbon and oxygen content, due to the different oxida-

tive treatment. Although both samples were able to generate

hydrogen from water, the less oxidized one (GO1) was more

efficient than the graphene sample with higher oxygen con-

tent (GO2). Using the less oxidized sample, they studied the

activity of mixtures of graphene oxide with tris(2,2-bipyridyl)

ruthenium(II), acridine orange hydrochloride, pyronin, meth-

ylene blue, thionin acetate, cis(diisothiocyanato)-bis(2,20-bipyridyl-4,40-dicarboxylate), ruthenium(II) bis (tetrabutyl

ammonium), copper phthalocyanine. The most efficient ma-

terial prepared was [Ru (bipy)3]2þ-graphene oxide complex,

able to produce hydrogen under sunlight illumination

(6 mmol g�1 h�1).

Recently, Lavorato et al. [80] started from chitosan pyroly-

sis to produce nitrogen-doped graphene photocatalysts which

were tested for hydrogen evolution under monochromatic UV

and visible light excitation from water/methanol solutions.

The novel nitrogen-doped graphene semiconductor showed

high efficiency for the photocatalytic generation of hydrogen

from water-methanol mixtures with similar efficiency using

UV or visible light.

The composition of two semiconductors possessing

different energy levels for their corresponding conduction and

valence bands, is another method to use visible light for

photocatalytic applications. When a large band gap semi-

conductor is coupled with a small band gap semiconductor

with a more negative conduction band level, photogenerated

electrons can be injected from the small band gap semi-

conductor to that with the large band gap. Thus, this tech-

nique provides amore efficient charge separation, an increase

in the life time of the charge carriers and an enhanced inter-

facial charge transfer to adsorbed substrates. Coupled semi-

conductors including TiO2/CdS, TiO2/SnO2, TiO2/ZnO and

TiO2/WO3 have been used intensively for organic photo-

degradation [81,82], but in the recent years they received

particular attention for their potential application in hydrogen

generation from water splitting [83]. It has been reported that

coupling CdS (band gap 2.4 eV) with TiO2 (band gap 3.2 eV)

hydrogen could be produced since the conduction band of

TiO2 is more negative than hydrogen production level

ðEH2=H2OÞ. In CdS/TiO2 the photogenerated electrons in CdS are

transferred into the TiO2 particles while the holes remain in

the CdS particle. This not only helps for charge separation by

isolating electrons and holes in two distinct particles but at

the same time, allows the extension of the photoresponse of

the photocatalyst in the visible region. CdS photocorrosion

can be prevented adding Na2S solution. Other composite

systems recently studied for the water splitting reaction are

ZnO/ZnS/CdS [13], ZnO/CdS [84], CdS/ZnS [85], InP/CdS [86]. In

general, the coupled system exhibit higher efficiency of the

photocatalytic process. Domen and co-workers [87] reported

that a NiO-loaded SrTiO3 powder was capable of decomposing

pure water as well as water vapor into H2 and O2 under UV

irradiation. The activity of the photocatalyst was increased

considerably by a pretreatment inH2 and using a concentrated

NaOH solution for the photocatalytic reaction. The photo-

catalytic activity of SrTiO3 was also greatly improved by using

a modified preparation method or a suitable concentration of

metal cations doping such as La3þ and Ga3þ.

Photocatalytic hydrogen generation in batchsystems

The first example of hydrogen generation fromwater splitting

upon visible light irradiation was reported only a decade ago

by Sayama et al. [88] and it was inspired by natural photo-

synthesis. A mixture of two different photocatalysts, Pt/WO3

for oxygen evolution and Pt/SrTiO3 (CreTa-doped) for

hydrogen generation using visible light irradiation

(l > 420 nm) and IO3�/I� redox couple as an electron mediator

were employed in a batch system. The overall water splitting

proceeded with hydrogen generation from water reduction

and I� oxidation to IO3� over PteTiO2-anatase, and IO3

� reduc-

tion to I� and water oxidation to O2 over TiO2-rutile. An

interesting aspect of this system was that H2 gas was evolved

only over the PteTiO2-anatasewhile O2 was produced over the

TiO2-rutile photocatalyst only.

Many others Z-scheme systems using reversible redox

mediators have been investigated to carry out the overall

water splitting under visible light irradiation. Kozlova et al.

[89] experimented the use of platinized titania as catalyst and

inorganic compounds as sacrificial agents for the generation

of H2 and O2 in a dual-step process. In particular the redox

couple Ce(III)/Ce(IV) was used for a cyclic transfer of electron

that allowed the generation of the two gases in different steps.

The consecutive generation of hydrogen and oxygen started

with the initial production of oxygen from water oxidation

and at the same time the reduction of Ce(IV) ions to the Ce(III)

form. The reaction proceeded until all Ce(IV) ions were con-

verted into Ce(III). In a second step, Pt/TiO2 catalyst was used

to carry out H2 formation and the concomitantly Ce(III)

oxidation to Ce(IV). Some drawbacks, such as the different

generation rate of the two gases, the catalyst concentration

and the formation of cerium salts during irradiation restricted

the possibility to use the dual-system, and they could be

overcome adding sulfuric acid in the reaction. This compound

promoting the formation of negatively charged complexes

was adsorbed on the semiconductor surface, increasing the

interactions between the metal ions and catalyst particles.

The efficiency of the use of Ce(IV) ions as electron accep-

tors for oxygen generation using different semiconductor

oxides is reported also in other works [90,91]. Bamwenda et al.

[90], reported the study of the photocatalytic decomposition of

water for the generation of oxygen with TiO2 rutile, TiO2

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 2 4 7e7 2 6 1 7253

anatase,WO3 and CeO2. The photocatalytic process was based

on the use of Ce(IV) and Fe(III) ions as electron acceptors,

selected both for their stability in an aqueous medium and for

their ability to be adsorbed on the catalyst surface, undergoing

reversible reduction by the photogenerated electrons. The

electron transfer from TiO2, WO3 and CeO2 to Ce(IV) and Fe(III)

is also thermodynamically favored because the standard

reduction potentials of the conduction band of the semi-

conductor have negative values, higher than the standard

reduction potentials of the Ce(IV)/Ce(III) and Fe (III)/Fe(II) pairs.

The photoreduction reaction may also be followed by a sub-

sequent oxidation of the reducedmetal ions as a result of their

interaction with the photogenerated holes. In Fe2(SO4)3aqueous solutions the photocatalytic efficiency was:

WO3 > TiO2 rutile > TiO2 anatase, while in suspensions con-

taining FeCl3, the TiO2 rutile led to better results compared to

that of WO3. A drawback of the process was represented by

the fact that the addition of Fe2(SO4)3 in aqueous suspensions

of TiO2 and CeO2 caused intermolecular aggregation. This

phenomenon was less pronounced when FeCl3 was used as

source of ions Fe(III), and negligible in suspension containing

WO3. In order to improve oxygen generation, the authors

propose the use of ion Ce(IV). This cation acts as effective

scavenger of photogenerated electrons and does not cause

intermolecular aggregation. The obtained results using Ce(IV)

showed higher yields of oxygen by using TiO2 as the

photocatalyst.

It should be noted that the catalytic efficiency of the water

splitting reaction was higher when rutile titania was used in

comparison to the anatase form, probably due to the different

light adsorption ability and to the different surface properties.

The rutile form generally have a higher concentration of hy-

droxyl groups on their surface compared to samples of

anatase. Such functional group play an important role in

determining the photocatalytic activity, since it acts as traps

for the photogenerated holes, forming hydroxyl radicals and

thus reducing the recombination reaction between electrons

and holes [92,93]. Furthermore, differences in the mobility of

the charges and in the speed of migration of the electrons

from themetal oxide particles to the 3d Fe(III) orbitals or to the

4f Ce(IV) orbitals, although the band gap values are almost

identical for both phases (2.8e3 eV). These factors seem to

favor the photocatalytic activity of the rutile phase with

respect to the anatase phase. Using the metallic pair Ce(IV)/

Ce(III) higher yields of oxygen in the initial stage of the process

were obtained compared to those obtained with Fe(III)/Fe(II),

while on long-term an opposite situation occurred. The dif-

ferences in initial reactivity recorded using the two pairs of

metals could be explained by considering a greater trapping of

electrons by the Ce(IV), related to differences in charge den-

sity, electron affinity, and redox potential of the pairs of the

metal ions. Thermodynamically, the standard potential of the

redox couple Ce(IV)/Ce (III), equal to þ1.44 V, has a positive

value higher than that of Fe(III)/Fe(II) of þ0.68 V. This resulted

into a greater driving force of the reduction of Ce(IV) than that

Fe(III) in redox systems. The rapid capture of electrons on the

surface of the semiconductor molecules in systems contain-

ing Ce(IV) causes an effective separation between the charge

with consequent lengthening of the life time of the holes

involved in the oxidation reaction of water, leading

consequently to higher yields of oxygen. For long-term pho-

tocatalytic runs, the higher yields observed using Fe(III)

compared to those with Ce(IV) could instead be due to the

ability of the ions Fe(II) to move away easily from the photo-

catalyst particles in suspension, preventing the recombina-

tion between electrons and holes.

The authors also report the study of the pH influence. The

best yields have been obtained working at pH values between

2 and 3 and the photogeneration of oxygen decreases with

increasing pH. This is in agreement with those reported by

Weber et al. [94], 1978, according to which the speed of pho-

tocatalytic production of oxygen decreases with increasing

pH.

The high efficiency of the photocatalytic process based on

the use of rutile titania and iron ions for producing oxygen

was also reported in the work of Ohno et al. [95] and it was

attributed to the preferential adsorption of Fe(III) ions with

respect to the Fe(II) ions on the surface of the photocatalyst.

The adsorption was favored in particular by using FeCl3rather than Fe2(SO4)3, suggesting a possible influence of the

anionic species on the interactions between the metal ions

and the semiconductor particles. Consideringwhat previously

exposed, it should be evidenced that the use ofmetal pairs like

Ce(IV)/Ce(III) and Fe(III)/Fe(II) is very effective for the water

splitting for oxygen generation, but still presents disadvan-

tages for the hydrogen production: the optimum operating

conditions for the production of the two gases are in fact

different.

Very recently, Maeda et al. [96] demonstrated the possi-

bility to obtain stoichiometric water splitting using BaZ-

rO3eBaTaO2N in combination with either PtOx/WO3 or rutile

TiO2. In presence of Fe(III)/Fe(II) the photocatalysts BaZ-

rO3eBaTaO2N and PtOx/WO3 led to the simultaneous genera-

tion of H2 and O2, although the performance was much lower

than that recorded with an IO3�/I� redox couple.

Photocatalytic hydrogen generation inmembrane systems

The development of photocatalytic applications requires an

efficient design of photocatalytic reactors for industrial-scale

use and commercial applications. Several aspects such as

type of irradiation source (natural or artificial), light source

position (immersed or external), catalyst (slurry or immobi-

lized on a support) play an important role in the photocatalytic

efficiency of the process [97,98]. In a photocatalytic reactor the

catalyst can be either suspended in aqueous medium or

immobilized on support materials, e.g. cellulose fibers, glass,

ceramic membrane [99e101]. In a slurry reactor, the catalyst

particles are suspended in the fluid phase. Some advantages

of this system include the fairly uniform catalyst distribution,

minimum catalyst fouling effects and practically no mass

transfer limitations. However, in this type of configuration

light scattering can occur, lowering the efficiency of the

treatment process. In addition, a post-process of catalyst

separation is needed to isolate the catalyst particles from the

reaction medium.

Immobilized photocatalytic reactors allow the continuous

use of the photocatalyst, eliminating the problem of the

Fig. 3 e Diagram of water splitting by a Z-scheme system using Au/CeO2 as photocatalyst for oxygen generation, Au/TiO2 for

hydrogen generation and Fe3D/Fe2D as redox couple (elaborated from Marino et al. [108]).

SO3- SO3

- SO3-

SO3-

SO3- SO3

-

H+

H+

H+

H+

H+H+ H+ H+

H+ H+H+ H+

Fig. 4 e Sulfonic groups of Nafion contribute to the proton

exchange activity (elaborated from Sørensen [109]).

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 2 4 7e7 2 6 17254

catalyst post-separation. Drawbacks of the use of this kind of

reactors are the possibility of mass transfer limitations,

catalyst fouling or wash-out, low surface to volume ratio and

significant pressure drop.

Slurry photocatalytic systems usually show largest pho-

tocatalytic activity when compared to immobilized photo-

catalytic reactors [102,103]. Increasing attention has being

taken to recently developed photocatalytic membrane re-

actors that could allow the selective separation of the prod-

uct(s) and/or intermediates from the reactive environment

[104e107]. The use of hybrid systems in which photo-

catalysis is coupled with a membrane module could have a

significant impact in the near future to design processes for

energy production, synthetic reactions and pollutant

abatement.

Most of the photocatalytic membrane reactors (PMRs)

experimented for hydrogen generation from water spitting

mimic the Z-scheme mechanism used by green plants for the

natural photosynthesis. Z-scheme basically includes a dual-

photocatalyst system for water oxidation (which lead to O2

formation) and for water reduction (with H2 generation). In

addition, a redox couple is used to regenerate the catalyst

transporting the charge between the two compartments of the

reactor. A membrane is required to separate the aqueous

suspensions containing the H2- and the O2-photocatalyst and

to shuffling electrons via the electron redox mediator (Fig. 3).

Generally a Nafion membrane, modified to enable the ions

transport, is used.

Titania and ceria samples, modified with gold for the

visible light adsorption, were used for H2 and O2 generation,

respectively, employing Fe(III)/Fe(II) as electrolyte [108].

Although preliminary tests in which Fe2þ was present in the

Au/TiO2 cell and Fe3þ in the Au/CeO2 compartment, led to

oxygen generation over both photocatalysts, when only Fe3þ

salt was added to the Au/CeO2 chamber, hydrogen and oxygen

were generated simultaneously in the cell containing Au/TiO2

and Au/CeO2 respectively.

Nafion is a copolymer of tetrafluoroethylene and per-

fluoro(4-methyl-3,6-dioxa-7-octene-1-sulfonyl-fluoride).

Nafion polymer is a thermoplastic resin that can be produced

into different shapes such as beads, film, and tubing. The

perfluorinated composition of the copolymer imparts chemi-

cal and thermal stability rarely available with nonfluorinated

polymers. The pendant sulfonyl fluoride groups are chemi-

cally changed to sulfonic acid [109] leading to a ionic func-

tionality. For example, by simple immersion in an acidic

solution, the film can be saturated with Hþ ions (Fig. 4).

The unique functional properties of Nafion polymer have

enabled a broad range of applications. Many efforts have been

made to incorporate semiconductor nanoparticles in a Nafion

film [110e113]. Deposition of Pt metal on one side of the film is

one strategy used to reduce Hþ into H2 [110].

The use of twin reactors shows several advantages in

comparison to the Z-scheme in a batch reactor: the backward

reaction of hydrogen and oxygen is minimized, the purifica-

tion process which allows the use of pure generated hydrogen

is not required and the safety issue of H2eO2 explosion is

overcame for a future industrial process.

The method of splitting water to produce hydrogen and

oxygen in separated sides was also studied in the Z-scheme

proposed by Fujihara et al. [95]. Unlike the other systems

described above, this device combines two photocatalytic

e-

Fig. 6 e Polymer membrane electrode assembly (MEA)

consisting of a TiO2 photoanode, a Pt cathode, and a proton

exchange membrane (PEM) used by Seger et al.

Photoanode: TiO2 (3.0 mg cmL2) loaded Toray paper;

photocathode: Pt particles dispersed on Toray carbon

paper (elaborated from [114]).

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 2 4 7e7 2 6 1 7255

reactions using TiO2 powders suspended in two different so-

lutions, and using a two-compartment cell equipped with Pt

electrode and a proton exchange membrane (Fig. 5).

When both compartments were irradiated by UV light, O2

was produced on rutile TiO2 particles suspended in a solution

containing Fe(III)/Fe(II), while H2 was generated on the TiO2

anatase modified with Pt in a solution containing Br2/Br�. In

the proposed system, Fe(II) ions formed in one compartment

as the result of reduction of the Fe(III) ions promoted by the

photogenerated electrons, were oxidized electrochemically

through an external circuit based on the release of Br-ion,

which in turn was photocatalytically oxidized to Br2. The

electrical neutrality and the pH of the solution in each

compartment were maintained constant through a proton

exchange membrane which allowed the proton transfer from

one side to the other of the system. Gases generation reached

stable values after 100 min (1.3 and 2.8 mmol h�1 for O2 and H2

respectively). Problems related to the deposition of iron

compounds on the Pt electrode which led to its contamina-

tion, were observed using this system and reduced gas gen-

eration rate during the time.

Seger et al. [114] reported a TiO2-Nafion-Pt assembly

membrane (Fig. 6) system using methanol as sacrificial agent

and UV light irradiation, without any external bias. TiO2

photocatalyst provided the methanol oxidation while, at the

same time, Hþ ions, driven to the Pt particle surface through

the Nafionmembrane, were reduced leading to a H2 evolution

rate of 69 mL h�1 cm�2.

A novel two-compartments reactor in which the advan-

tages of the Z-scheme and the H-type reactor were combined,

was developed by Wu et al. [115]. In this system, powders of

WO3 and Pt/SrTiO3:Rh were employed to generate oxygen and

hydrogen, respectively. A cation exchange membrane placed

between the two semi-cells of the reactor, allowed the trans-

fer of both protons and Fe(II)/Fe(III) ions in solution.

Later, in another study, hydrogen and oxygen were pro-

duced using Pt/SrTiO3:Rh and BiVO4 by Yu et al. [116] in a twin-

reactor system using a Nafion membrane (Fig. 7).

Fig. 5 e Schematization of the photocatalytic reactor used

by Fujihara et al. for the separated hydrogen and oxygen

generation (elaborated from [95]).

In order to exchange the functional groups of Nafion from

Hþ to Fe3þ or Fe2þ form, the commercial membrane was

immersed in FeCl3 or FeCl2 solutions for 18 h. Pt/SrTiO3:Rhwas

prepared by the solid-state fusion procedure starting from

SrCO3, TiO2 and Rh2O3 and used as hydrogen photocatalyst,

and BiVO4 was prepared by the aqueous synthesis method

and used for oxygen generation. H2 and O2 were produced in

stoichiometric ratio for the entire photocatalytic run, leading

to a formation rate of 0.65 and 0.32 mol h�1 g�1 of photo-

catalyst. The influence of the initial concentration of the iron

species was also investigated and the obtained data revealed

that hydrogen generation increased with the increasing of the

iron concentration although not proportional to the gas pro-

duction. Starting from an 2mMFe(II) aqueous solution (pH 2.4)

in which 0.1 g L�1 of Pt/SrTiO3:Rh was suspended, a hydrogen

evolution rate of 0.74 mol h�1 g�1 was obtained in the first

three hours of the photocatalytic run.

Various research groups [117e120] propose the use of

devices based on thin films of TiO2. These innovative

H

hνν hνν

BiVOPt/SrTiO :Rh

Fig. 7 e Membrane twin reactor experimented by Yu et al.

(elaborated from [116]).

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 2 4 7e7 2 6 17256

photocatalysts have been studied also for environmental and

energy processes, as the photocatalytic purification of waste

water and air from toxic compounds, the development of solar

cells and photochemical applications [121]. Thin films of TiO2

synthesized by solegel process [122], although they showed

high photocatalytic reactivity, had poor stability and me-

chanical durability. On the contrary, preparing the catalyst

with the sputtering procedure, promising results for the con-

struction of high density coatings of large surface areas with a

strong adhesion and uniform, was obtained. Anpo et al. [123]

proposed the use of a photocatalyst consisting of a Ti sheet

(thickness of 59 mm) deposited on a side of a thin TiO2

responsive to visible light, prepared by working at a temper-

ature of 873.80 K. This device was mounted in a glass

container and separates two aqueous solutions. The side of

the thin TiO2 film was immersed in an 1.0 N NaOH aqueous

solution, while the Pt side was immersed in a solution of 1.0 N

H2SO4. Irradiating the device with visible light (l > 450 nm)

lead to the production of stoichiometric separate hydrogen

and oxygenwith a good linearity to the irradiation time. A thin

film device, prepared by Radio Frequency Magnetron Sput-

tering (RF-MS) procedure, consisting of a 50 mm Ti foil and a

titania film one side, and with Pt on the other side, mounted

on an H-type glass container separating two aqueous solu-

tions was studied by Kitano et al. [119]. A chemical bias be-

tween the two compartments was created using a 1.0 M NaOH

aqueous solution in the titania side and a 0.5MH2SO4 aqueous

solution in the Pt side (Fig. 8).

The two gases were generated in stoichiometric amount,

reaching 0.088 and 0.043 mmol h�1 for H2 and O2 respectively,

under visible light irradiation (l � 450 nm).

In the work by Selli et al. [120], the water splitting reaction

was studied in a two Plexiglas compartments that allowed the

separate production of hydrogen and oxygen on the two

different side of a photoactive Ti electrode, obtained by

deposition of thin TiO2 films on discs of pure Ti. One side of

the cell was irradiated with a UV lamp through a window (side

Fig. 8 e Schematization of the H-type photocatalytic reactor

for the separated H2 and O2 production used by Kitano et al.

(elaborated from [119]).

of 4 cm). Themodified Ti disc with surface area of 10 cm2, was

mounted between the two compartments of the cell, with the

TiO2 layer exposed to the irradiation source. A cation-

exchange membrane was mounted over the Ti disc sepa-

rating the two compartments of the system. Also in this case,

in order to generate a small electrochemical current which

ensured the passage of electrons two aqueous solutions of

NaOH and H2SO4 were used (Fig. 9).

The results showed that oxygen was the main gas gener-

ated on the titania irradiated side, acting as photoanode, while

hydrogen was the main gas produced from water reduction,

on the Pt/Ti side. The photocatalytic runs were carried out

using Ti electrodes with different chemical-physical charac-

teristics, evidencing in particular a promising efficiency for

the electrode prepared with the magnetron sputtering depo-

sition of Pt on TiO2 operating at 600 �C which maintained high

stability working continuously for over 180 h. Under 300 nm

wavelength light irradiation, the H2 production rate was

w100molmin�1 and that of O2w 26molmin�1. It’s interesting

to note that the production of the two gases increased when

the normality of the NaOH and H2SO4 was increased at values

higher than 1.0 N, while their production was negligible when

the two compartments were filled with pure water.

Recently, an innovative membrane photo-system was

proposed by Tsydenov et al. [124] based on a porous polymeric

membrane modified with Pt and titania to allow hydrogen

generation from ethanol dehydrogenation. A polytetrafluoro-

ethylene filter modified with polypropylene, thanks to its high

physical-chemical stability under light irradiation, was used

as support (Fig. 10).

The influence of several aspects such as the optimum

titania and metal loading, the mechanical resistance, perme-

ability and morphology of the membrane, were analyzed in

order to improve the performance of the catalytic membrane

system on the hydrogen generation. Photocatalytic tests were

carried out using aqueous ethanol solution under inert at-

mosphere to produce H2. The reaction led to the formation of

methane, ethane, carbon dioxide and hydrogen as gaseous

Fig. 9 e Schematization of the photocatalytic membrane

system used by Selli et al. (elaborated from [120]).

Fig. 10 e Schematization of the process of photocatalytic

hydrogen and oxygen generation from water splitting

based on the membrane system proposed by Tsydenov

et al. (elaborated from [124]).

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 2 4 7e7 2 6 1 7257

products and acetic acid and acetaldehyde as liquid by-

products. A quantum efficiency of about 13% was registered

using an UV irradiation source (l < 345 nm, irradiance

187mW cm�2) and gave a H2 evolution rate of about 15ml h�1.

Dual-layer photo-electrode preparation was investigated

by Liao et al. [125] for the visible light water splitting into

hydrogen and oxygen. The photocatalyst, consisting in a layer

of WO3 and a layer of TiO2 deposited on a Pt coated foil, was

tested in an H-type reactor and in photovoltammetry experi-

ments, showing good catalytic properties in the visible light

adsorption. Photocatalytic runs were carried out using UV and

visible light evidencing a better performance compared to that

of the titania electrode.

Marschall et al. [126] built amembrane system (Fig. 11) with

two chambers for pure hydrogen and oxygen collection in

which the membrane separated and at the same time sup-

ported a carbon coated Degussa TiO2 P25 photoanode and a Pt

cathode. No sacrificial agent were used in this system. Several

membrane types, such as commercial Nafion, FKE (Fumatech)

membranes, handmade sulfonated polyethersulfone (sPES)

flat sheet [127] and Spes/mesoporous-Si-MCM-41-

Fig. 11 e Schematic representation of membrane-

electrode-photocatalyst-assembly proposed by Marschall

et al. (elaborated from [126]).

nanoparticles [127,128] were experimented in the photo-

catalytic system.

The novel composite sPES40membrane containing 0.5 wt%

sulfonated mesoporous Si-MCM-41 nanoparticles (SN) led to

promising results for hydrogen generation (82.4 mmol h�1 for

titania P25 in water, and 104.5 mmol h�1 using diluted HCl)

that, in addition to the absence of sacrificial agents, makes

this system a promising membrane system.

Very recently Kudo et al. [129] experimented a visible light

irradiated membrane system using tris(2,2-bipyridyl) cobalt

(II), ([Co(bpy)3]3þ/2þ) electron mediator. A membrane having

10 mm pore size was used to separate two aqueous solutions

containing Ru/SrTiO3:Rh and BiVO4 catalysts for hydrogen and

oxygen evolution, respectively (Fig. 12). After 3 h irradiation, a

hydrogen production rate of 110 mL h�1 was detected in the Ru/

SrTiO3:Rh-side, while in the compartment containing BiVO4

both oxygen (43 mL h�1) and hydrogen (4.3 mL h�1) were pro-

duced, probably because H2 from the Ru/SrTiO3:Rh-compart-

ment pass through the membrane. The obtained results

showed also that [Co(bpy)3]3þ/2þ play a key role for gases for-

mation as long as it canmove from one side to the other of the

system.

The systems and experimental devices described in this

paragraph show a key role of a membrane for the photo-

generation of separate streams of hydrogen and oxygen

from water. The most interesting ones are that using: i) an H-

type reactor, ii) a Z-scheme process, iii) visible light irradia-

tion, iv) no use of sacrificial reagents, v) irradiation of only

one-side of the cell, vi) a continuous process. Which is the

most convenient photocatalyst configuration (immobilized in

thin films or suspended powders in slurry systems) is amatter

of study so as the preparation of novel photocatalysts able to

split water with high conversion under visible light (sunlight)

irradiation. However, membrane reactors (or hybrid processes

using also membranes) in this field for photocatalytic

hydrogen (and oxygen) generation are still at early stage of

development. Too much efforts must be done to built a robust

Fig. 12 e Membrane system used by Kudo et al. for the

separated hydrogen and oxygen generation from water

splitting using sunlight irradiation (elaborated from [129]).

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 2 4 7e7 2 6 17258

system able to work with high overall conversion efficiency.

Use of a suitable membrane so as its configuration represent

two key steps in this challenging research topic for harvesting

solar energy to produce hydrogen fuel from a renewable

source as water.

Conclusions

Hydrogen generation via water splitting is a powerful green

process. Several methods such as noble metal loading, ion

doping, dye-sensitization, composite photocatalysts, have

been developed to enhance the photocatalytic activity of

semiconductors under visible light irradiation, giving prom-

ising results. They shouldmodify the band gap of thematerial

for harnessing a greater portion of visible light and also reduce

the charge recombination mechanism.

Use of sacrificial agents (hole scavenger and electron

donor) could improve the hydrogen production minimizing

the backward reaction of recombination of the photo-

generated charges.

Membrane photo-reactors are effective to attain sustain-

able hydrogen production by preventing the backward reac-

tion between H2 and O2, and they allow to obtain pure

hydrogen in a single step, without further purification before

its utilization. Although many efforts are necessary to

improve the performance of the process, membrane photo-

systems represent a novel interesting way to make a cost-

effective and environmentally sustainable water splitting

process for hydrogen production.

Appendix A. Supplementary data

Supplementary data related to this article can be found online

at http://dx.doi.org/10.1016/j.ijhydene.2014.02.174.

r e f e r e n c e s

[1] Trimm DL, Onsan ZI. Onboard fuel conversion forhydrogen-fuel-cell-driven vehicles. Catal Rev Sci Eng2001;43(1e2):31e84.

[2] Chaubey R, Sahu S, James OO, Maity S. A review ondevelopment of industrial processes and emergingtechniques for production of hydrogen from renewable andsustainable sources. Renew Sustain Energy Rev2013;23:443e62.

[3] Pakshirajan K, Mal J. Biohydrogen production using nativecarbon monoxide converting anaerobic microbialconsortium predominantly Petrobacter sp. Int J HydrogenEnergy 2013;38(36):16020e8. http://dx.doi.org/10.1016/j.ijhydene.2013.09.129.

[4] Liu C, Lei Z, Yang Y, Zhang Z. Preliminary trial ondegradation of waste activated sludge and simultaneoushydrogen production in a newly-developed solarphotocatalytic reactor with AgX/TiO2-coated glass tubes.Water Res 2013;47(14):4986e92.

[5] Liao C-H, Huang C-W, Wu JCS. Hydrogen production fromsemiconductor-based photocatalysis via water splitting.Catalysts 2012;2(4):490e516.

[6] Holladay JD, Hu J, King DL, Wang Y. An overview ofhydrogen production technologies. Catal Today2009;139(4):244e60.

[7] Lee JS. Photocatalytic water splitting under visible light withparticulate semiconductor catalysts. Catal Surv Asia2005;9(4):217e27.

[8] Maeda K, Domen K. Photocatalytic properties of RuO2-loaded b-Ge3N4 for overall water splitting. J Phys Chem C2007;111(12):4749e55.

[9] Inoue Y. Photocatalytic water splitting by RuO2-loadedmetal oxides and nitrides with d0- and d10 -relatedelectronic configurations. Energy & Environ Sci2009;2(4):364e86.

[10] Kudo A, Miseki Y. Heterogeneous photocatalyst materialsfor water splitting. Chem Soc Rev 2009;38(1):253e78.

[11] Youngblood WJ, Lee S-HA, Maeda K, Mallouk TE. Visiblelight water splitting using dye-sensitized oxidesemiconductors. Acc Chem Res 2009;42(12):1966e73.

[12] Maeda K, Domen K. Photocatalytic water splitting: recentprogress and future challenges. J Phys Chem Lett2010;1(18):2655e61.

[13] Chen X, Shen S, Guo L, Mao SS. Semiconductor-basedphotocatalytic hydrogen generation. Chem Rev2010;110(11):6503e70.

[14] Maeda K. Photocatalytic water splitting usingsemiconductor particles: history and recentdevelopments. J Photochem Photobiol C Photochem Rev2011;12(4):237e68.

[15] Kudo A. Z-scheme photocatalyst system for water splittingunder visible light irradiation. MRS Bull 2011;36(1):32e8.

[16] Abe R. Development of a new system for photocatalyticwater splitting into H2 and O2 under visible light irradiation.Bull Chem Soc Jpn 2011;84(10):1000e30.

[17] Wang Y, Wang X, Antonietti M. Polymeric graphitic carbonnitride as a heterogeneous organocatalyst: fromphotochemistry to multipurpose catalysis to sustainablechemistry. Angew Chem Int Ed 2012;51(1):68e89.

[18] Osterloh FE. Inorganic nanostructures forphotoelectrochemical and photocatalytic water splitting.Chem Soc Rev 2013;42(6):2294e320.

[19] Maeda K. (Oxy)nitrides with d0-electronic configuration asphotocatalysts and photoanodes that operate under a widerange of visible light for overall water splitting. Phys ChemChem Phys 2013;15:10537e48.

[20] Molinari R, Caruso A, Argurio P, Poerio T. Degradation of thedrugs gemfibrozil and tamoxifen in pressurized and de-pressurized membrane photoreactors using suspendedpolycrystalline TiO2 as catalyst. J Membr Sci2008;319(1e2):54e63.

[21] Molinari R, Gallo S, Argurio P. Metal ions removal fromwastewater or washing water from contaminated soil byultrafiltrationecomplexation. Water Res2004;38(3):593e600.

[22] Molinari R, Argurio P, Pirillo F. Comparison betweenstagnant sandwich and supported liquid membranes incopper(II) removal from aqueous solutions: flux, stabilityand model elaboration. J Membr Sci 2005;256(1e2):158e68.

[23] Molinari R, Argurio P, Romeo L. Studies on interactionsbetween membranes (RO and NF) and pollutants (SiO2, NO3

�,Mnþþ and humic acid) in water. Desalination2001;138(1e3):271e81.

[24] Molinari R, Argurio P, Lavorato C. Review on reduction andpartial oxidation of organics in photocatalytic (membrane)reactors. Curr Org Chem 2013;17(21):2516e37.

[25] Palmisano L, Sclafani A. Thermodynamics and kinetics forheterogeneous photocatalytic processes. In: Schiavello M,editor. Heterogeneous photocatalysis. West Sassex, UK:John Wiley & Sons; 1997.

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 2 4 7e7 2 6 1 7259

[26] Hoffmann MR, Martin ST, Choi W, Bahnemann DW.Environmental applications of semiconductorphotocatalysis. Chem Rev 1995;95(1):69e96.

[27] Litter M. Heterogeneous photocatalysis: transition metalions in photocatalytic systems. Appl Catal B: Environ1999;23(2e3):89e114.

[28] Ward MD, White JR, Bard AJ. Electrochemical investigationof the energetics of particulate titanium dioxidephotocatalysts. The methyl viologen-acetate system. J AmChem Soc 1983;105(1):27e31.

[29] Ni M, Leung MKH, Leung DYC, Sumathy K. A review andrecent developments in photocatalytic water-splitting usingTiO2 for hydrogen production. Renew Sustain Energy Rev2007;11(3):401e25.

[30] Fujishima A, Honda K. Electrochemical photolysis of waterat a semiconductor electrode. Nature 1972;238:37e8.

[31] Shi W, Song S, Zhang H. Hydrothermal synthetic strategiesof inorganic semiconducting nanostructures. Chem Soc Rev2013;42(13):5714e43.

[32] Daghrir R, Drogui P, Robert D. Modified TiO2 forenvironmental photocatalytic applications: a review. IndEng Chem Res 2013;52(10):3581e99.

[33] Ibhadon AO, Fitzpatrick P. Heterogeneous photocatalysis:recent advances and applications. Catalysts2013;3(1):189e218.

[34] Qu Y, Duan X. Progress, challenge and perspective ofheterogeneous photocatalysts. Chem Soc Rev2013;(42):2568e80.

[35] Sarina S, Waclawika ER, Zhu H. Photocatalysis on supportedgold and silver nanoparticles under ultraviolet and visiblelight irradiation. Green Chem 2013;15(7):1814e33.

[36] Diebold U. The surface science of titanium dioxide. Surf SciReports 2003;48(5e8):53e229.

[37] Di Paola A, Cufalo G, Addamo M, Bellardita M,Campostrini R, Ischia M, et al. Photocatalytic activity ofnanocristalline TiO2 (brookite, rutile and brookite-based)powders prepared by thermohydrolysis of TiCl4 in aqueouschloride solutions. Colloids Surf Physicochem Eng Asp2008;317(1):366e76.

[38] Howe RF. Recent developments in photocatalysis. DevChem Eng Miner Process 1998;6(1e2):55e84.

[39] Fischer CH, Lillie J, Weller H, Katsikas L, Henglein A.Photochemistry of colloidal semiconductors. 29.Fractionation of CdS sols of small particles by exclusionchromatography. Berichte Bunsen Phys Chemistry1989;93(1):61e4.

[40] Beydoun D, Amal R, Low G, McEvoy S. Role of nanoparticlesin photocatalysis. J Nanoparticle Res 1999;1(4):439e58.

[41] Fox MA, Dulay MT. Heterogeneous photocatalysis. ChemRev 1993;93(1):341e57.

[42] Trillas M, Pujol M, Domenech X. Phenol photodegradationover titanium dioxide. J Chem Technol Biotechnol1992;55(1):85e90.

[43] Navaladian S, Janet CM, Viswanathan B, Viswanath RP. Onthe possible treatment procedures for organiccontaminants. Photoelectrochem Photobiol Environ Energyfuel 2007;37:1e51. Research signpost.

[44] Wu CH. Comparison of azo dye decolorization efficiencyusing UV/single semiconductor and UV/coupledsemiconductor systems. Chemosphere 2004;57(7):601e8.

[45] Lai CW, Hamid SBA, Sreekantan S. A novel solar drivenphotocatalyst: well-aligned anodic WO3 nanotubes. Int JPhotoenergy; 2013, 6 pages, http://dx.doi.org/10.1155/2013/745301 [Article ID 745301].

[46] Oliveira HG, Fitzmorris BC, Longo C, Zhang JZ.Photoelectrochemical and photocatalytic properties of TiO2,WO3 and WO3-TiO2 porous films in the photodegradation of

rhodamine 6G in aqueous solution. Sci Adv Mater2012;4(5e6):673e80.

[47] Wang F, Di Valentin C, Pacchioni G. Rational band gapengineering of WO3 photocatalyst for visible light watersplitting. ChemCatChem 2012;4(4):476e8.

[48] Santato C, Odziemkowski M, Ulmann M, Augustynski J.Crystallographically oriented mesoporous WO3 films:synthesis, characterization, and applications. J Am ChemSoc 2001;123(43):10639e49.

[49] Khan SB, Faisal M, Rahman MM, Akhtar K, Asiri AM,Khan A, et al. Effect of particle size on the photocatalyticactivity and sensing properties of CeO2 nanoparticles. Int JElectrochem Sci 2013;8:7284e97.

[50] Arul NS, Mangalaraj D, Kim TW, Chen PC, Ponpandian N,Meena P, et al. Synthesis of CeO2 nanorods with improvedphotocatalytic activity: comparison between precipitationand hydrothermal process. J Mater Sci Mater Electron2013;24(5):1644e50.

[51] Zhou XD, Huebner W, Anderson HU. Processing ofnanometer-scale CeO2 particles. Chem Mater2003;15(2):378e82.

[52] Sathyamurthy S, Leonard KJ, Dabestani RT,Paranthaman MP. Reverse micellar synthesis of ceriumoxide nanoparticles. Nanotechnology 2005;16(9):1960e4.

[53] Campelo JM, Luna D, Luque R, Marinas JM, Romero AA.Sustainable preparation of supported metal nanoparticlesand their applications in catalysis. ChemSusChem2009;2(1):18e45.

[54] Z. Jin, M. Xiao, Z. Bao, P. Wang, and J. Wang, “A generalapproach to mesoporous metal oxide microspheres loadedwith Noble metal nanoparticles”, Angew Chem Int Ed, vol.51, no. 26, pp. 6406e6410.

[55] Mandal S, Roy D, Chaudhari RV, Sastry M. Pt and Pdnanoparticles immobilized on amine-functionalizedzeolite: excellent catalysts for hydrogenation and heckreactions. Chem Mater 2004;16(19):3714e24.

[56] Xiong Z, Ma J, Ng WJ, Waite TD, Zhao XS. Silver-modifiedmesoporous TiO2 photocatalyst for water purification.Water Res 2011;45(5):2095e103.

[57] Lim MA, Lee YW, Han SW, Park I. Novel fabrication methodof diverse one-dimensional Pt/ZnO hybrid nanostructuresand its sensor application. Nanotechnology 2011;22(3).http://dx.doi.org/10.1088/0957-4484/22/3/035601 [Article ID035601].

[58] Periasamy C, Chakrabarti P. Time-dependent degradation ofPt/ZnO nanoneedle rectifying contact based piezoelectricnanogenerator. J Appl Phys 2011;109. http://dx.doi.org/10.1063/1.3553862 [Article ID 054306].

[59] Hara M, Hitoki G, Takata T, Kondo JN, Kobayashi H,Domen K. TaON and Ta3N5 as new visible light drivenphotocatalysts. Catal Today 2003;78(1e4):555e60.

[60] Wu G, Chen T, Zhou G, Zong X, Li C. H2 production with lowCO selectivity from photocatalytic reforming of glucose onmetal/TiO2 catalysts. Sci China Ser B 2008;51(2):97e100.

[61] Gurunathan K. Photocatalytic hydrogen production usingtransition metal ions-doped g-Bi2O3 semiconductorparticles. Int J Hydrogen Energy 2004;29(9):933e40.

[62] Mizukoshi Y, Sato K, Konno TJ, Masahashi N. Dependenceof photocatalytic activities upon the structures of Au/Pdbimetallic nanoparticles immobilized on TiO2 surface. ApplCatal B: Environ 2010;94(3e4):248e53.

[63] Bond GC, Sermon PA, Webb G, Buchanan DA, Wells PB.Hydrogenation over supported gold catalysts. J Chem SocChem Commun; 1973. 444b-445.

[64] Haruta M, Kobayashi T, Sano H, Yamada N. Novel goldcatalysts for the oxidation of carbon monoxide at atemperature far below 0 �C. Chem Lett 1987;16(2):405e8.

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 2 4 7e7 2 6 17260

[65] Silva CG, Juarez R, Marino T, Molinari R, Garcıa H. Influenceof excitation wavelength (UV or visible light) on thephotocatalytic activity of titania containing goldnanoparticles for the generation of hydrogen or oxygenfrom water. J Am Chem Soc 2011;133(3):595e602.

[66] Haruta M. Gold as a novel catalyst in the 21st century:preparation, working mechanism and applications. GoldBull 2004;37(1e2):27e36.

[67] Hutchings GJ. New directions in gold catalysis. Gold Bull2004;37(1e2):3e11.

[68] Primo A, Marino T, Corma A, Molinari R, Garcıa H. Efficientvisible-light photocatalytic water splitting by minuteamounts of gold supported on nanoparticulate CeO2

obtained by a biopolymer templating method. J Am ChemSoc 2011;133(18):6930e3.

[69] Kang Q, Lu QZ, Liu SH, Yang LX, Wen LF, Luo SL, et al. Aternary hybrid CdS/Pt-TiO2 nanotube structure forphotoelectrocatalytic bactericidal effects on Escherichia Coli.Biomaterials 2010;31(12):3317e26.

[70] Abe R, Sayama K, Sugihara H. Development of newphotocatalytic water splitting into H2 and O2 using twodifferent semiconductor photocatalysts and a shuttle redoxmediator IO3

�/I�. J Phys Chem B 2005;109(33):16052e61.[71] Shet S, Ahn KS, Deutsch T, Wang HL, Nuggehalli R, Yan YF,

et al. Influence of gas ambient on the synthesis of co-dopedZnO:(Al,N) films for photoelectrochemical water splitting. JPower Sources 2010;195(17):5801e5.

[72] Kudo A, Sekizawa M. Photocatalytic H2 evolution undervisible light irradiation on Ni-doped ZnS photocatalyst.Chem Commun 2000;2000(15):1371e2.

[73] Kudo A, Kato H. Effect of lanthanide-doping into NaTaO3

photocatalysts for efficient water splitting. Chem Phys Lett2000;331(5e6):373e7.

[74] O’Regan B, Graetzel M. A low-cost, high-efficiency solar cellbased on dye-sensitized colloidal TiO2 films. Nature1991;353:737e40.

[75] Law KY. Organic photoconductive materials: recent trendsand developments. Chem Rev 1993;93(1):449e86.

[76] Dhanalakshmi KB, Latha S, Anandan S, Maruthamuthu P.Dye sensitized hydrogen evolution from water. Int JHydrogen Energy 2001;26(7):669e74.

[77] Borgarello E, Kiwi J, Pelizzetti E, Visca M, Gratzel M.Photochemical cleavage of water by photocatalysis. Nature1981;289:158e60.

[78] Nada AA, Hamed HA, Barakat MH, Mohamed NR,Veziroglu TN. Enhancement of photocatalytic hydrogenproduction rate using photosensitized TiO2/RuO2-MV2þ. IntJ Hydrogen Energy 2008;33(13):3264e9.

[79] Latorre-Sanchez M, Lavorato C, Puche M, Fornes V,Molinari R, Garcia H. Visible-light photocatalytic hydrogengeneration by using dye-sensitized graphene oxide as aphotocatalyst. Chem A Eur J 2012;18(52):16774e83.

[80] Lavorato C, Primo A, Molinari R, Garcia H. “N-dopedgraphene derived from biomass as visible lightphotocatalyst for hydrogen generation from water-methanol mixtures”. Chem A Eur J, 2014;20(1):187e94.

[81] Lin CF, Wu CH, Onn ZN. Degradation of 4-chlorophenol inTiO2, WO3, SnO2, TiO2/WO3 and TiO2/SnO2 systems. JHazard Mater 2008;154(1e3):1033e9.

[82] Wang X, Liu M, Chen Q, Zhang K, Chen J, Wang M, et al.Synthesis of CdS/CNTs photocatalysts and study ofhydrogen production by photocatalytic water splitting. Int JHydrogen Energy 2013;38(29):13091e6.

[83] Baniasadi E, Dincer I, Naterer GF. Hybrid photocatalyticwater splitting for an expanded range of the solar spectrumwith cadmium sulfide and zinc sulfide catalysts. Appl CatalA Gen 2013;455(30):25e31.

[84] Wang X, Liu G, Chen ZG, Li F. Highly efficient H2 evolutionover ZnO-ZnS-CdS heterostructures from an aqueoussolution containing SO3

2� and S2� ions. J Mater Res2010;25(1):39e44.

[85] Grimes CA, Varghese OK, Ranjan S. Hydrogen generation bywater splitting. In: Light, water, hydrogen e the solargeneration of hydrogen by water photoelectrolysis. US:Springer; 2008, ISBN 978-0-387-33198-0. pp. 35e113. http://dx.doi.org/10.1007/978-0-387-68238-9.

[86] Liu B, Bando Y, Liu L, Zhao J, Masanori M, Jiang X, et al.Solid-solution semiconductor nanowires in pseudobinarysystems. Nano Lett 2013;13(1):85e90.

[87] Domen K, Kudo A, Onishi T, Kosugi N, Kuroda H.Photocatalytic decomposition of water into hydrogen andoxygen over nickel(II) oxide-strontium titanate (SrTiO3)powder. 1. Structure of the catalyst. J Phys Chem B1986;90(2):292e5.

[88] Sayama K, Mukasa K, Abe R, Abe Y, Arakawa H.Stoichiometric water splitting into H2 and O2 using amixture of two different photocatalysts and an IO3

�/I�

shuttle redox mediator under visible light irradiation. ChemCommun 2001;2001(23):2416e7.

[89] Kozlova AE, Tatyana P, Korobkina P, Vorontsov VA. Overallwater splitting over Pt/TiO2 catalyst with Ce3þ/Ce4þ shuttlecharge transfer system. Int J Hydrogen Energy2009;34(1):138e46.

[90] Bamwenda GR, Arakawa H. The photoinduced evolution ofO2 and H2 from a WO3 suspension in the presence of Ce4þ/Ce3þ. Sol Energy Mater Sol Cells 2001;70(1):1e14.

[91] Bamwenda GR, Uesigi T, Abe Y, Sayama K, Arakawa H. Thephotocatalytic oxidation of water to O2 over pure CeO2, WO3

and TiO2 using Fe3þ and Ce4þ as electron acceptors. ApplCatal A: General 2001;205(1e2):117e28.

[92] Park DR, Zhang J, Ikeue K, Yamashita H, Anpo M.Photocatalytic oxidation of ethylene to CO2 and H2O onultrafine powdered TiO2 photocatalysts in the presence ofO2 and H2O. J Catal 1999;185(1):114e9.

[93] Lin J, Yu JC, Lo D, Lam SK. Photocatalytic activity of rutileTi1�xSnxO2 solid solutions. J Catal 1999;183(2):368e72.

[94] Weber J, Samec Z, Marecek V. The effect of anion adsorptionon the kinetics of the Fe3þ/Fe2þ reaction on Pt and Auelectrodes in HClO4. J Electroanal Chem InterfacialElectrochem 1978;89(2):271e88.

[95] Fujihara K, Ohno T, Matsumara M. Splitting of water byelectrochemical combination of two photocatalyticreactions on TiO2 particles. J Chem Soc Faraday Trans1998;94(24):3705e9.

[96] Maeda K, Lu D, Domen K. Solar-driven Z-scheme watersplitting using modified BaZrO3-BaTaO2N solid solutions asphotocatalysts. ACS Catal 2013;3(5):1026e33.

[97] Xing Z, Zong X, Pan J, Wang L. On the engineering part ofsolar hydrogen production from water splitting:photoreactor design. Chem Eng Sci 2013;104:125e46.

[98] Maeda K. Z-Scheme water splitting using two differentsemiconductor photocatalysts. ACS Catal2013;3(7):1486e503.

[99] A. Chica, “Zeolites: promised materials for the sustainableproduction of hydrogen”, ISRN Chem Eng, Volume 2013,Article ID 907425, http://dx.doi.org/10.1155/2013/907425.

[100] Zeng J, Liu S, Cai J, Zhang L. TiO2 immobilized in cellulosematrix for photocatalytic degradation of phenol under weakUV light irradiation. J Phys Chem C 2010;114(17):7806e11.

[101] Lazar MA, Varghese S, Nair SS. Photocatalytic watertreatment by titanium dioxide: recent updates. Catalysts2012;2(4):572e601.

[102] Umar M, Abdul Aziz H. Photocatalytic degradation oforganic pollutants in water. In: Rashed MN, editor. Organic

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 2 4 7e7 2 6 1 7261

pollutants e monitoring, risk and treatment, ISBN 978-953-51-0948-8. http://dx.doi.org/10.5772/53699 [Chapter 8].

[103] de Lasa H, Serrano B, Salaices M. Novel photocatalyticreactors for water and air treatment. Photocatalytic ReactEng; 2005:17e47. http://dx.doi.org/10.1007/0-387-27591-6_2.

[104] Bellobono IR, Morazzoni F, Tozzi PM. Photocatalyticmembrane modules for drinking water purification indomestic and community appliances. Int J Photoenergy2005;7(3):109e13.

[105] Molinari R, Palmisano L, Drioli E, Schiavello M. Studies onvarious reactor configurations for coupling photocatalysisand membrane processes in water purification. J Membr Sci2002;206(1e2):399e415.

[106] Teplyakov VV, Gassanova LG, Sostina EG, Slepova EV,Modigell M, Netrusov AI. Lab-scale bioreactor integratedwith active membrane system for hydrogen production:experience and prospects. Int J Hydrogen Energy2002;27(11e12):1149e55.

[107] Rhoden SLNH, Mettee HD, Linkous CA. Development of ahydrogen evolving photocatalytic membrane. EnergyProcedia 2009;29:522e31.

[108] Marino T, Primo A, Corma A, Molinari R, Garcıa H.Photocatalytic overall water splitting by combining goldnanoparticles supported on TiO2 and CeO2. In: 2ndEuropean Symposium on Photocatalysis, Bordeaux (France);29the30th September 2011. p. 76.

[109] Sørensen B. Hydrogen and fuel cells: emerging technologiesand applications. Academic Press; 2005, ISBN 978-0-12-655281-2.

[110] Mau AWH, Huang CB, Kakuta N, Bard AJ, Campion A,Fox MA, et al. Hydrogen photoproduction by nafion/cadmium sulfide/platinum films in water/sulfide ionsolutions. J Am Chem Soc 1984;106(22):6537e42.

[111] Kakuta N, White JM, Campion A, Bard AJ, Fox MA,Webber SE. Surface analysis of semiconductor-incorporatedpolymer systems. 1. Nafion and cadmium sulfide-Nafion. JPhys Chem 1985;89(1):48e52.

[112] Willner I, Steinberger-Willner B. “Solar hydrogenproduction through photobiological, photochemical andphotoelectrochemical assemblies. Int J Hydrogen Energy1988;13(10):593e604.

[113] Maeda K, Teramura K, Lu D, Takata T, Saito N, Inoue Y, et al.Photocatalyst releasing hydrogen from water. Nature2006;440:295.

[114] Seger B, Kamat PV. Fuel cell geared in reverse:photocatalytic hydrogen production using a TiO2/Nafion/Ptmembrane assembly with no applied bias. J Phys Chem C2009;113(43):18946e52.

[115] Lo C-C, Huang C-W, Liao C-H, Wu JCS. Novel twin reactor forseparate evolution of hydrogen and oxygen inphotocatalytic water splitting. Int J Hydrogen Energy2010;35(4):1523e9.

[116] Yu S-C, Huang C-W, Liao C-H, Wu JCS, Chang S-T, Chen K-H. A novel membrane reactor for separating hydrogen andoxygen in photocatalytic water splitting. J Membr Sci2011;382(1e2):291e9.

[117] Zhu J, Zach M. Nanostructured materials for photocatalytichydrogen production. Curr Opin Colloid Interface Sci2009;14(4):260e9.

[118] Matsuoka M, Kitano M, Takeuchi M, Tsujimaru K, Anpo M,Thomas J. Photocatalysis for new energy production.Recent advances in photocatalytic water splittingreactions for hydrogen production. Catal Today2007;122(1e2):51e61.

[119] Kitano M, Tsujimaru K, Anpo M. Decomposition of water inthe separate evolution of hydrogen and oxygen using visiblelight-responsive TiO2 thin film photocatalysts: effect of thework function of the substrates on the yield of the reaction.Appl Catal A Gen 2006;314(2):179e83.

[120] Selli E, Chiarello GL, Quartarone E, Mustarelli P, Rossetti I,Forni L. A photocatalytic water splitting device for separatehydrogen and oxygen evolution. Chem Commun2007;2007(47):5022e4.

[121] Takeuchi M, Dohshi S, Eura T, Anpo M. Preparation oftitanium-silicon binary oxide thin photocatalysts by anionized cluster beam deposition method. Theirphotocatalytic activity and photoinduced super-hydrophilicity. J Phys Chem B 2003;107(51):14278e82.

[122] Miyauchi M, Nakajima A, Watanabe T, Hashimoto K.Photoinduced hydrophilic conversion of TiO2/WO3 layeredthin films. Chem Mater 2002;14(11):4714e20.

[123] Anpo M. Preparation, characterization and reactivities ofhighly functional titanium oxide based photocatalysts ableto operate under UV-visible light irradiation: approaches inrealizing high efficiency in the use of visible light. BullChem Soc Jpn 2004;77(8):1427e42.

[124] Tsydenov DE, Parmon VN, Vorontsov AV. Toward thedesign of asymmetric photocatalytic membranes forhydrogen production: preparation of TiO2-basedmembranes and their properties. Int J Hydrogen Energy2012;37(15):11046e60.

[125] Liao C-H, Huang C-W, Wu JCS. Novel dual-layerphotoelectrode prepared by RF magnetron sputtering forphotocatalytic water splitting”. Int J Hydrogen Energy2012;37(16):11632e9.

[126] Marschall R, Klaysom C, Mukherji A, Wark M, Lu GQM,Wang L. “Composite proton-conducting polymermembranes for clean hydrogen production with solar lightin a simple photoelectrochemical compartment cell. Int JHydrogen Energy 2012;37(5):4012e7.

[127] Klaysom C, Marschall R, Wang LZ, Ladewig BP, Lu GQM.Synthesis of composite ion-exchange membranes and theirelectrochemical properties for desalination applications. JMater Chem 2010;20(22):4669e74.

[128] Marschall R, Bannat I, Feldhoff A, Wang LZ, Lu QGM,Wark M. Nanoparticles of mesoporous SO3H-functionalizedSi-MCM-41 with superior proton conductivity. Small2009;5(7):854e9.

[129] Sasaki Y, Kato H, Kudo A. [Co(bpy)3]3þ/2þ and [Co(phen)3]

3þ/

2þ electron mediators for overall water splitting undersunlight irradiation using Z-scheme photocatalyst system. JAm Chem Soc 2013;135(14):5441e9.