photochemical splitting of water for hydrogen production

Review

Photochemical splitting of water for hydrogen productionby photocatalysis: A review

Adel A. Ismail a,b,c,n, Detlef W. Bahnemann c

a Advanced Materials Department, Central Metallurgical R&D Institute, CMRDI, P.O. Box 87, Helwan 11421, Cairo, Egyptb Advanced Materials and NanoResearch Center, Najran University, P.O. Box: 1988, Najran 11001, Saudi Arabiac Institute fr Technische Chemie, Leibniz Universitt Hannover, Callinstrasse 3, D-30167 Hannover, Germany

a r t i c l e i n f o

Article history:Received 26 January 2014Received in revised form21 April 2014Accepted 27 April 2014Available online 2 June 2014

Keywords:Water splittingHydrogen productionPhotocatalystsUV and visible illumination

a b s t r a c t

Hydrogen production from water using a catalyst and solar energy is an ideal future fuel source. The searchfor suitable semiconductors as photocatalysts for water splitting into molecular hydrogen and oxygen hasbeen considered to be an urgent subject for our daily life. In this review, we aim to focus on the researchefforts that have been made so far for H2 generation from water splitting by UV and visible light-drivenphotocatalysis. A number of synthetic modication methods for adapting the electronic structure to enhancethe charge separation in the photocatalyst materials are discussed. Sacricial reagents and electronmediators for the overall water splitting are also reviewed. The quantum efciency of photocatalystmaterials upon visible and UV illumination will be reviewed, summarized and discussed.

& 2014 Elsevier B.V. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 861.1. Basic principles of photocatalytic hydrogen generation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 861.2. Main processes of photocatalytic hydrogen generation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 871.3. Evaluation of photocatalytic water splitting systems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

2. Photocatalysts for water splitting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 872.1. UV photocatalysts for water splitting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

2.1.1. TiO2 and titanates as photocatalysts for H2 production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 872.1.2. Nb- and Ta-based oxides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 882.1.3. W-, Mo-, and V-based oxides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 892.1.4. d10 Metal oxide photocatalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 892.1.5. f0 Metal oxide photocatalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 892.1.6. Nonoxide photocatalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

2.2. Visible light active photocatalysts for H2 production. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 892.2.1. Efcient separation of photogenerated charge carriers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 902.2.2. Cocatalyst loading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 902.2.3. Z-scheme photocatalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 902.2.4. Metal suldes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 922.2.5. GaN:ZnO solid solution photocatalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 922.2.6. Overall water splitting employing dye-sensitized solar cells to harvest visible light . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 952.2.7. Electron mediators for overall water splitting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96

3. Main challenges and opportunities. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 974. Conclusions and outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97

Contents lists available at ScienceDirect

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

Solar Energy Materials & Solar Cells

http://dx.doi.org/10.1016/j.solmat.2014.04.0370927-0248/& 2014 Elsevier B.V. All rights reserved.

n Corresponding author.E-mail address: [email protected] (A.A. Ismail).

Solar Energy Materials & Solar Cells 128 (2014) 85101

Acknowledgment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98

1. Introduction

Photocatalytic water splitting into H2 and O2 using semiconduct-ing catalysts has received much attention due to the potential ofthis technology, as well as the great economic and environmentalinterest for the production of the clean fuel H2 from water usingsolar energy. Fujishima and Honda demonstrated the potential ofTiO2 semiconductor materials to split water into H2 and O2 [1,2].Their work triggered the development of semiconductor photo-catalysis for a wide range of environmental and energy applications[1,2]. The photocatalytic activity of conventional metal-oxide photo-catalysts for the overall water splitting is known to be heavilydependent on the crystallinity and the particle size of the material,as determined by the preparation conditions [35]. During the past40 years, various photocatalyst materials have been developed tosplit water into H2 and O2 under UV and visible light illumination.The direct splitting of water using a particulate photocatalyst wouldbe a good way to produce clean and recyclable H2 on a largescale [69]. A number of photocatalysts have been proposed andachieved high quantum efciencies under UV illumination. Atpresent, there is a lack of suitable materials with sufciently bandgap positions for overall water splitting, and the stability necessaryfor practical applications. In general, efcient photocatalytic materi-als contain either transition-metal cations with a d0 electronicconguration (e.g., Ta5 , Ti4 , Zr4 , Nb5 , Ta5 , W6 , andMo6) or typical metal cations with d10 electronic conguration(e.g., In3 and Sn4 , Ga3 , Ge4 , Sb5) as principal cationcomponents, the empty d or sp orbitals of which form the bottomof the respective conduction bands [69]. The tops of the valencebands of metal oxide photocatalysts with d0- or d10-metal cationsusually consist of O2p orbitals, which are located at about 3 eV orhigher versus normal hydrogen electrode (NHE) and, as such,produce a band gap too wide to absorb visible light [9]. Also,(oxy) nitrides containing d0 transition-metal cations, such as Ta3N5,TaON, and LaTiO2N, are potential photocatalytic materials to achievewater splitting [10]. The highest quantum efciencies (QEs) havebeen reported for NiO-modied La/KTaO3 (QE56%) using waterunder UV light [11], QE90% has been obtained for H2 generationemploying ZnS as photocatalysts in the presence of aqueous Na2S/Na2SO3 as sacricial reductant under UV illumination [12] and forCr/Rh-modied GaN/ZnO (QE2.5%), in pure water under visiblelight illumination [1315]. So far, no material capable of catalyzing

reaction (Eq. (1)) with visible light and a QE larger than 10% hasbeen found [15]. Therefore, investigating the physical factors thatgovern the photocatalytic activity for H2 production is an importantand indispensable issue in the development of highly active photo-catalysts. While a number of visible-light-driven photocatalystshave been proposed as potential candidates for the overall watersplitting, a satisfactory material has yet to be devised [1619]. At apower level of 1000 W/m2 the solar energy incident on the earth'ssurface by far exceeds all human energy needs [20,21]. Photovoltaicand electrochemical solar cells that convert solar energy intoelectricity can reach up to 5577% efciency [2225] but remainuneconomical because of high fabrication costs, insufcient lightabsorption [26], and inefcient charge transfer [22]. In a processthat mimics photosynthesis; solar energy can also be used toconvert water into H2 and O2, the fuels of a H2-based energyeconomy (Eq. (1))

H2O-1=2 O2H2g; G 237 kJ=mol1:3 eV=e; min 1100 nm1

The development of a photocatalytic system that functionsefciently under visible light, representing almost half of theavailable solar energy on the surface of the earth, is thereforeessential for the practical utilization of solar energy. In this review,we aim to summarize the research efforts having been made so far,for H2 generation from water splitting employing UV and visiblelight-driven photocatalysts in combination with the view oninspiring new ideas to tackle this important challenge. A numberof synthetic and modication techniques for adjusting the bandstructure of the photocatalyst to harvest visible light and toimprove the charge separation in photocatalysis are discussed.Sacricial reagents and electron mediators for the overall watersplitting are also reviewed.

1.1. Basic principles of photocatalytic hydrogen generation

Fig. 1a shows a schematic diagram of water splitting into H2 andO2 over photocatalysts. Photocatalysis on semiconductor particlesinvolves three main steps: (i) absorption of photons with energiesexceeding the semiconductor bandgap, leading to the generation ofelectron (e) and hole (h) pairs in the semiconductor particles;(ii) charge separation followed by migration of these photogener-ated carriers in the semiconductor particles; (iii) surface chemical

Fig. 1. Schematic illustration of water splitting over semiconductor photocatalysts.Source: Ref. [103].

A.A. Ismail, D.W. Bahnemann / Solar Energy Materials & Solar Cells 128 (2014) 8510186

reactions between these carriers with various compounds (e.g.,H2O); electrons and holes may also recombine with each otherwithout participating in any chemical reactions [3,7]. When aphotocatalyst is used for water splitting, the energetic position ofthe bottom of the conduction band must be more negative than thereduction potential of water to produce H2, and that of the top ofthe valence band must be more positive than the oxidationpotential of water to produce O2, as shown in Fig. 1b. Furthermore,the photocatalyst must be stable in aqueous solutions underphotoirradiation. Scaife noted in 1980 that it is intrinsically difcultto develop an oxide semiconductor photocatalyst that has both,a sufciently negative conduction band for H2 production and asufciently narrow band gap (i.e.,o3.0 eV) for visible light absorp-tion because of the highly positive position of the valence band(at ca. 3.0V vs. NHE) formed by the O2p orbitals [27]. To facilitatewater oxidation, the potential of the valence band edge must exceedthe oxidation potential of water of 1.23 V vs NHE at pH0(0.82 V at pH7). Based on these parameters, a theoreticalsemiconductor band-gap energy of 1.23 eV is required to drivethe water-splitting reaction according to Eq. (1). The smallest bandgap achieved so far in a functional catalyst is 2.30 eV in NiO/RuO2Ni:InTaO4 [28,29].

Semiconductors with smaller band gaps or lower at-bandpotentials require a bias voltage or external redox reagents to drivethe reaction. Alternatively, two or more small band-gap semicon-ductors can be combined to drive water oxidation/reduction pro-cesses separately via multiphoton processes. It is well known thatat-band potentials strongly depend on ion absorption (protonationof surface hydroxyl groups), crystallographic orientation of theexposed surface, surface defects, and surface oxidation processes[30,31]. These factors are rarely considered in the preparation andtesting of photochemical water-splitting catalysts [3234].

1.2. Main processes of photocatalytic hydrogen generation

The individual processes involved in the photocatalytic gen-eration of H2 are illustrated in Fig. 1. They include light absorptionof the semiconductor photocatalyst, generation of excited chargecarriers (electrons and holes), recombination of these chargecarriers, separation of the excited charge carriers, migration ofelectrons and holes, charge carrier trapping, and transfer of thecharge carriers to water or other molecules [3,7]. All of theseprocesses affect the nal generation of H2 from the semiconductorphotocatalyst system. The total amount of H2 generated is deter-mined by the amount of excited electrons at the water/photo-catalyst interface capable of reducing water. After the electron/hole pairs are created, charge recombination and separation/migration processes are two important competitive processesinside the semiconductor photocatalyst that largely affect theefciency of the photocatalytic reaction for water splitting [35].Charge recombination reduces the number of e/h pairs byemitting light or generating phonons. This includes both, surfaceand bulk recombination, and is classied as a deactivation process,and it is ineffective for water splitting. Efcient charge separationand fast charge carrier transport, avoiding any bulk/surface chargerecombination, are thus fundamentally important for the photo-catalytic H2 generation through water splitting. The reaction ofphotogenerated H2 and O2 to form H2O on the photocatalystsurface is normally called surface backreaction (SBR). There aretwo main approaches to suppress the SBR: one involves theaddition of sacricial reagents into the photocatalytic reactionenvironment and the second creates a separation of the photo-active sites on the surface of the photocatalysts. In general,sacricial reagents acting as electron donors or acceptors, respec-tively, drive the reaction into alternative pathways as they arereduced or oxidized, respectively. Taking into consideration the

basic mechanism and the individual processes of photocatalyticwater splitting, there are two keys for the development of asuitable high-efciency semiconductor for the visible-light-driven photocatalytic splitting of water into H2 and/or O2: (1) aphotocatalyst should have a sufciently narrow band gap(1.23 eVoEgo3.0 eV) to both harvest visible light and possessthe correct band structure; and (2) photoinduced charges in thephotocatalyst should be separated efciently in order to avoidbulk/surface electron/hole recombination. In addition, both chargecarriers must migrate to the photocatalyst surface for H2 and/or O2evolution at the respective photocatalytic active sites [36]. It isimportant that economical, highly efcient photocatalytic systemsfor light-to-H2 energy conversion, in which aqueous solutionscontaining sacricial reagents can be used to minimize the back-ward reaction of H2 and O2 to water on the surface of photo-catalysts, can be constructed.

1.3. Evaluation of photocatalytic water splitting systems

The rate of gas (O2 and H2) evolutionwith units such as mol h1is used to enable a measurable comparison between differentphotocatalysts under similar experimental conditions. The quantumyield, as an extension from the overall quantum yield dened in ahomogeneous photochemical system, becomes important and accep-table to evaluate the photocatalytic activity for water splitting.The overall quantum yield is dened for H2 and O2 formation byEqs. (2) and (3), respectively [37]. The quantum yield is estimated tobe smaller than the overall quantum yield because the number ofabsorbed photons is usually smaller than the number of photonsavailable in the incident light. Activities of photochemical water-splitting catalysts are usually assessed by the rates of evolved gases[mol/h] per catalyst amount [g] under the specied irradiationconditions. From the measured evolution rate [H2], the apparentQE2[H2]/I of the catalyst can be calculated using the known photonux [mol/s] incident on the reaction mixture.Overall quantum yield%

2 Number of evolved H2 moleculesNumber of incident photons

100for H2 evolution

2

Overall quantum yield% 4 Number of evolvedO2 molecules

Number of incident photons 100for O2 evolution

3

2. Photocatalysts for water splitting

2.1. UV photocatalysts for water splitting

A wide range of semiconducting materials have been devel-oped as photocatalysts under UV irradiation. On the basis of theirelectronic conguration properties, these UV-active photocatalystscan be typically classied into four groups: (1) d0 metal (Ti4 ,Zr4 , Nb5 , Ta5 , W6 , and Mo6) oxide photocatalysts, (2) d10

metal (In3 , Ga3 , Ge4 , Sn4 , and Sb5) oxide photocatalysts,(3) f0 metal (Ce4) oxide photocatalysts, and (4) a small group ofnonoxide photocatalysts.

2.1.1. TiO2 and titanates as photocatalysts for H2 productionTiO2 was the rst material described as a photochemical water-

splitting catalyst (Fig. 2). It crystallizes in three structure types:Rutile, Anatase, and Brookite. All modications contain TiO6octahedra that are interconnected via two (Rutile), three (Broo-kite), or four (Anatase) common edges and via shared corners, and

A.A. Ismail, D.W. Bahnemann / Solar Energy Materials & Solar Cells 128 (2014) 85101 87

as a result, the band gaps (3.0 eV for Rutile and 3.15 eV forAnatase) differ slightly [38]. In their 1971/72 papers, Fujishimaand Honda described an electrochemical cell consisting of a n-typeTiO2 (Rutile) anode and a Pt black cathode [1,2]. When the cell wasirradiated with UV light (o415 nm) from a 500 W Xe lamp, O2evolution was observed at the anode with a current owing to thePt counter electrode. Based on the current, a photoelectrochemicalefciency of approx. 10% was estimated.

Under UV irradiation, PtTiO2 was found to catalyze the waterreduction with iodide as the sacricial electron donor, whereasTiO2 was observed to be the water oxidation catalyst itself in thepresence of IO3 as the electron acceptor. After both catalysts werecombined, H2 and O2 were formed stoichiometrically from a basic(pH11) solution, with iodide serving as a redox shuttle. A similarsystem was realized with PtTiO2 and PtWO3, giving QE4%upon irradiation with UV light [39]. Other recent efforts havesought to improve the optical response of TiO2-based catalysts viadoping with C or N, and S [4043]. Pt-modied catalysts evolve O2from aqueous AgNO3 as the sacricial electron acceptor and tracesof H2 from aqueous methanol as the sacricial electron donorunder visible light [44]. Newly synthesized tailored anatase/brookite mixtures as well as pure brookite TiO2 nanorods werealso employed for H2 production. The results achieved so farrevealed that anatase/brookite mixtures and brookite nanorodsexhibit higher photocatalytic activity than anatase nanoparticlesand even higher efciencies than TiO2 P25 for the photocatalyticH2 evolution from aqueous methanol solution [45,46]. Further-more, mesoporous TiO2 nanostructures have been studied as aroute to minimize Pt loading on TiO2 photocatalysts for H2production. The results indicated that the amount of H2 evolvedon 0.2 wt% Pt/TiO2 calcined at 450 1C is three times higher thanthat evolved on Pt/TiO2-P25 and twelve times higher than thatevolved on Pt/TiO2 calcined at 350 1C [47]. Also, a series of TiO2and graphene sheet(GSs) composites were synthesized [48]. Thephotocatalytic activity of these samples was evaluated by H2evolution from water photo-splitting under UVvis illumination.On the other hand, when TiO2 is fused with metal oxides (SrO,PbO, etc.) metal titanates with intermediate band gaps can beobtained. Of these, SrTiO3 crystallizes in the Perovskite structuretype and has a band gap of 3.2 eV, slightly larger than that of TiO2.It was rst employed in 1976 as a photocatalyst in a water-splittingelectrochemical cell together with p-CdTe or p-GaP photocathodes[49]. An optimized version of this system was reported in 1977 byOhashi and co-workers to be the rst self-supported photoelec-trochemical cell [50] with a photon to-electron conversion ef-ciency of 0.0440.67%. In 1980, it was shown that NiO-modied

SrTiO3 powder splits water vapor stoichiometrically under UVirradiation, while SrTiO3 alone did not show any activity [51]. Theconditions of the NiO deposition had a strong effect on thecatalytic activity [52]. Abe et al. used Cr/Ta-doped Pt/SrTiO3together with PtWO3 in a two-particle catalyst system for overallwater splitting with visible light. Cr/Ta:SrTiO3Pt produced H2 andPtWO3 produced O2, with an iodide/iodate redox couple servingas the redox mediator between the two catalysts [53]. The effect ofmetal cocatalysts (Ru, Ir, Pd, Pt, Os, Re, Co) on the water-splittingactivity of SrTiO3 was also studied [54,55]. From pure water, H2and O2 were evolved stoichiometrically, with the activity decreas-ing in the order Rh4Ru4Re4Pt4 Ir4Pd4Os4Co. SrTiO3alone produced H2 but no O2. A series of layered Perovskites asphotochemical water-splitting catalysts was investigated [56,57].The NiO-modied catalyst is active for H2/O2 evolution with a QEup to 12% [5658]. By doping with BaO and upon addition of NaOHto the catalyst suspension, this activity could be further increasedto QE50% [56], only slightly below that of the best catalyst, i.e.,La-doped NaTaO3 with QE56% [59,60].

2.1.2. Nb- and Ta-based oxidesNb2O5 has a band gap of 3.4 eV and it is not active for pure

water splitting under UV irradiation. After modication with Pt asa cocatalyst, however, it can efciently produce H2 from aqueoussolutions containing methanol as an electron donor [61]. Meso-porous Nb2O5, demonstrated a photocatalytic activity 20 timeshigher for H2 evolution than a bulk Nb2O5 without any porosity.Intercalation of In2O3 into the mesoporous structure furtherincreased the photoactivity of mesoporous Nb2O5 by 2.7 times[62]. Besides Nb2O5, a large number of niobates can produce H2and O2 via water splitting upon UV irradiation. K4Nb6O17 wasdeveloped as the rst example of a niobate photocatalyst thatshowed high and stable activity for H2 evolution from aqueousmethanol solution without any assistance from other materialssuch as the noble metals [63,64]. There are various nanocompo-sites of SrNb2O6, Sr2Nb2O7, and Sr5Nb4O15 that exhibited efcientphotocatalytic activities for H2 and O2 production under UVirradiation [6567]. In particular, Sr2Nb2O7 with a highly donor-doped layered perovskite structure gave quantum yields as high as23%. The activity of Sr2Nb2O7 was further enhanced achievingquantum yields up to 32% [65]. Kudo and co-workers reportedBa5Nb4O15 with a layered perovskite structure for water splittingloaded with NiO cocatalysts and reported 17% quantum yield upon270 nm illumination [66,67]. Partial substitution of Nb5 withZn2 to obtain BaZn1/3Nb2/3O3 with a distorted perovskite struc-ture showed favorable photocatalytic activities under UV irradia-tion [68,69]. The photocatalytic activity of ZnNb2O6 was negligible,whereas NiO-loaded ZnNb2O6 revealed a high photoactivity [70].Chen and coworkers prepared ABi2Nb2O9 (ACa, Sr, Ba) perovs-kite photocatalysts for both H2 and O2 evolution from aqueoussolutions containing sacricial reagents [71]. The photocatalyticperformance decreased in the order of SrBi2Nb2O94BaBi2N-b2O94CaBi2Nb2O9. M2BiNbO7 (M In3 ,Ga3) with pyrochlorestructures were sensitive to UV irradiation and had the ability tosplit water [72]. Sabio et al. studied the photocatalytic watersplitting activities of suspended KCa2Nb3O10 nanoscale and bulkparticles and compared them using a kinetic model [73]. Theirmodel showed that the higher activity of the nanoniobate can berationalized by shortened transport paths for electrons and holesto the reactive surface, which reduce the lattice recombination.This effect apparently outweighs the increase in surface recombi-nation that is to be expected for the nanomaterial, due to itsincreased specic surface area.

Fig. 2. Schematic representation of a photoelectrochemical cell (PEC).Source: Reprinted with permission from Ref. [1].


2.1.3. W-, Mo-, and V-based oxidesPbWO4 incorporating a WO4 tetrahedron showed high and

stable photocatalytic activity for the overall splitting of water[74,75]. H2 and O2 was produced under UV irradiation when RuO2was loaded onto the metal oxide. The photocatalytic performancewas attributed to large dispersions in both the valence andconduction bands. PbMoO4 catalyzed the H2 evolution fromaqueous methanol solution. It also exhibited photocatalytic O2evolution from aqueous AgNO3 solution under UV irradiation withits O2 evolution activity being comparable to that observed forTiO2 [76]. Na2W4O13 and Bi2W2O9, with layered structures, wereactive for photocatalytic H2 and O2 evolution under UV irradiation[77,78]. However, Bi2MoO6 with a similar structure evolved onlyO2 from aqueous AgNO3 solution at a low rate [79]. A new crystalstructure of VO2, with a body centered-cubic structure (bcc) and alarge optical band gap of 2.7 eV, surprisingly showed excellentphotocatalytic activity for H2 production under UV irradiation [80].The bcc VO2 phase exhibited a high quantum efciency of 38.7%when synthesized as nanorods. Also, Bi2GaVO7 and Bi2YVO8 withtetragonal structures were studied. These two compoundsinitiated both, H2 and O2 evolution, from water under UV irradia-tion. This is in spite of the fact that both of them showed strongoptical absorption in the visible region (4420 nm) [81,82].

2.1.4. d10 Metal oxide photocatalystsVarious typical metal oxides (In3 , Ga3 , Ge4 , Sn4 , Sb5)

with d10 were shown to be effective photochemical water-splittingcatalysts under UV irradiation. Ni-loaded Ga2O3 appears to be oneof the promising photocatalysts for overall water splitting [83]. Itsphotocatalytic activity could be effectively improved by the additionof Ca, Cr, Zn, Sr, Ba, and Ta ions [84]. Zn ion doping remarkablyimproved the photocatalytic activity, with an apparent quantumyield for Ni/ZnGa2O3 of QE20%. By combining with Lu2O3, theresulting Zn-doped Lu2O3/Ga2O3 proved to be a novel compositephotocatalyst for stoichiometric water splitting under UV irradia-tion. NiO was loaded as the cocatalyst, the quantum yield at 320 nmamounted to 6.81%. Ga1.14In0.86O3 showed the highest photocataly-tic activity for H2 and O2 evolution [85]. In comparison, Y1.3In0.7O3,showed the highest photocatalytic activity for the overall watersplitting when combined with RuO2 as a promoter [86].

2.1.5. f0 Metal oxide photocatalystsCeO2 was reported to show a consistent activity towards O2

production in aqueous solutions containing Fe3 and Ce4 aselectron acceptors [87]. Sr2-doped CeO2 was an active photo-catalyst for overall water splitting when RuO2 was loaded as apromoter [88]. Ce3 supported zeolites showed higher photoca-talytic activity for water splitting [89]. H2 and O2 evolution wasobserved. Photoirradiation of Ce3 species generated electrons(Ce3 h-Ce4e) that were captured effectively by watermolecules resulting the formation of molecular hydrogen. Yuanet al. [90] reported that BaCeO3 produced H2 and O2 from aqueoussolutions containing CH3OH and AgNO3 as sacricial reagents,respectively.

2.1.6. Nonoxide photocatalystsZnS in SO32 solutions were used for H2 production under UV

irradiation exhibiting 90% quantum yield at 313 nm [91]. CdSenanoribbons in Na2S/Na2SO3 solution showed photocatalytic activ-ity for H2 evolution under both, UV and visible light illumination.The quantum efciency for this process has been reported to be0.09% [92]. The photocatalytic water-splitting activity of GaN wasfound to be strongly dependent on the crystallinity of the materialand on the cocatalyst employed [93]. Modication of well-crystallized GaN with Rh2yCryO3 nanoparticles as a cocatalyst

for H2 evolution resulted in the stable stoichiometric decomposi-tion of H2O into H2 and O2 under UV irradiation. RuO2 modica-tion, on the other hand, did not bring about appreciable H2 and O2evolution. However, Zn2 , Mg2 , and Be2 doping of GaN con-verted it into a remarkably active and stable photocatalyst. Again,the presence of RuO2 as a cocatalyst was required [94]. -Ge3N4was another effective nitride photocatalyst to show efcientactivity for splitting water into H2 and O2 when combined withRuO2 nanoparticles as reported by Domen's group [95,96]. Thephotocatalytic activity of RuO2-loaded -Ge3N4 was stronglydependent on the reaction conditions employed. The highestactivity was obtained when the reaction was carried out in 1 Maqueous H2SO4 solution. Moreover, treatment of as-prepared-Ge3N4 powder under high-pressure ammonia effectively increasedthe photocatalytic activity by up to 4 times. This was attributed toa decrease in the density of anion defects in the bulk and at thesurface [97]. A AgBr/SiO2 catalyst showed a stable and highphotocatalytic activity for H2 generation from CH3OH/H2O solutionunder UV irradiation. The high activity of this AgBr/SiO2 catalyst hasbeen related to photogenerated Ag species acting as sites for H2formation [98]. Very recently, metal-free polymeric photocatalystshave attracted much attention. Wang et al. demonstrated thatgraphitic carbon nitrides (g-C3N4) can be photocatalytically activefor H2 evolution from water under visible light irradiation [99].By introducing a mesoporous structure into polymeric C3N4, theefciency of H2 production from the photochemical reduction ofwater is increased by one order of magnitude and the quantumefciency (QE) is 1.4% [100,101].

2.2. Visible light active photocatalysts for H2 production

Although, more than 100 photocatalytic systems based onmetal oxides have been reported to be active for the watersplitting, most of them require ultraviolet (UV) light (o400nm) due to the large bandgap of these semiconductor materials[16101]. Since nearly half of the solar energy incident to theearth's surface lies in the visible region (400 nmoo800 nm)(see Fig. 3), it is essential to use visible light efciently to realize H2production on a large scale by photocatalytic water splitting.A maximum solar conversion efciency for photocatalytic watersplitting with a quantum efciency of 100% can be calculated usingthe standard solar spectrum. Even if all UV light up to 400 nmwere utilized, the solar conversion efciency would be only 2%,which is similar to the maximum conversion efciencies ofphotosynthesis in green plants under normal environmental

Fig. 3. Solar spectrum and maximum solar light conversion efciencies for watersplitting reaction with 100% of quantum efciency.Source: Ref. [103].


conditions (12%) [102]. However, if it was possible to utilizevisible light up to 600 nm, the efciency would drastically improveto 16%; a further extension up to 800 nm would give a conversionefciency of 32%. Therefore, achieving water splitting under visiblelight has been a challenging goal since the discovery of theHondaFujishima effect in 1972 [1,2]. Despite many years ofintensive efforts by researchers around the world, the rst repro-ducible demonstration of water splitting under visible lightirradiation was reported only a decade ago [11,39,103].

2.2.1. Efcient separation of photogenerated charge carriersWhile visible-light-driven photocatalysts with proper band

structures are currently being developed using some modicationtechnologies or band engineering approaches, the issue of photo-generated charge carrier separation is another key factor stronglyaffecting the efciency of the photocatalytic water-splitting pro-cess [3,103]. The energy band conguration of a semiconductorplays a signicant role in the absorption of light and in determin-ing the redox potentials [3]. In order to narrow the bandgap ofsemiconductors to extend the absorption of light into the visibleregion, three approaches have been widely used: (I) modicationof the VB, (II) adjustment of the CB, and (III) continuous modula-tion of the VB and/or CB [104] (Fig. 4).

2.2.2. Cocatalyst loadingTransition metals, especially noble metals, are widely used as

effective cocatalysts for photocatalytic readtions [54,55]. The

processes of charge transfer between cocatalyst and host photo-catalyst are described in Fig. 5 [7]. When the noble metal is loadedonto the surface of the photocatalyst, the photogenerated elec-trons migrating to the surface of the host photocatalyst areentrapped by the noble metal cocatalyst, because the Fermi energylevel of noble metal is always lower than that of the semiconduc-tor photocatalyst [46,47]. Meanwhile, the photogenerated holesstay at the host photocatalyst and also migrate to its surface.Subsequently, the separately localized electrons and holes will beinvolved as reducing and oxidizing agents, respectively. Theimportance of cocatalysts as follows: It improves the overallphotocatalytic activity of the water splitting and it acceleratesthe surface chemical reaction by inhibiting the backward reaction.To date, the most widely used co-catalysts are noble metals (Pt, Ru,Rh, Pd) and metal oxides (NiO, RuO2) [105]. Compared to theinvestigation of photocatalysts, efforts in developing co-catalystsare quite rare. However, the design and discovery of new types ofco-catalyst may play an equally important role in the developmentof photocatalysts themselves by nally enabling a high H2 evolu-tion efciency for practical applications. For example, a rhodiumchromium mixed oxide as co-catalyst led to coupled GaN:ZnOphotocatalysts with 2.5% quantum efciency for water splittingunder visible light irradiation [15]. Au as a cocatalyst wasimproved the photocatalytic water splitting over La-doped NaTaO3[106]. Besides noble metals and metal oxides, metal suldes havealso been explored as active cocatalysts. MoS2 as a co-catalyst hasbeen reported to be even more effective than Pt for CdS watersplitting systems [107].

2.2.3. Z-scheme photocatalystsThere are two main approaches for achieving water splitting

using visible light. One approach is to apply a two-step photo-excitation mechanism between two photocatalysts (a two-stepsystem; Fig. 6a) [11,39,53,78,108112]. This process was inspiredby natural photosynthesis in green plants and is known as thephotocatalytic Z-scheme. In two-step systems, the water splittingreaction is broken up into two stages: one for H2 evolution and theother for O2 evolution; these are combined by using a shuttleredox couple (Red/Ox) in the solution. At the H2 evolutionphotocatalyst, the photoexcited electrons reduce water to H2 whilethe holes in the valence band oxidize the reductant (Red) to itsoxidized form (Ox). This system lowers the energy required forphotocatalysis, allowing visible light to be utilized more efcientlythan in conventional water-splitting systems (Fig. 6b) [103].Visible-light responsive oxides such as WO3 can be used as O2photocatalysts if they can reduce the oxidant to a reductant.Indeed, different semiconductor materials can be used in theFig. 4. Three strategies to narrow the bandgap of semiconductor photocatalysts to

match the solar spectrum.Source: Ref. [3,104].

Fig. 5. Processes of charge transfer between host photocatalyst and cocatalyst.Source: Ref. [35].

Fig. 6. Schematic energy diagrams of photocatalytic water splitting systems:(a) two-step photoexcitation system and (b) conventional one-step system.Source: Ref. [103].


Z-scheme even if they do not satisfy all the stringent requirementsfor a one-step system. Z-scheme systems has ability to separatethe production of H2 and O2 by employing a separator. Thedisadvantage of Z-scheme systems. it requires the number ofphotons to be twice as large as the one-step system to achievewater splitting. The other approach is a single visible-lightresponsive photocatalyst, as illustrated in Fig. 6b. Because only afew semiconductor materials can absorb visible light and have asufciently high potential for water splitting, band engineering ofsemiconductors is required to nd out new photocatalysts thatsatisfy the following requirements for water splitting upon visiblelight illumination [103]: exhibit a narrow band gap, be stableunder photoirradiation, and have suitable conduction and valenceband levels for H2 and O2 production, respectively. Various bandengineering methods have been employed, including the intro-duction of midgap electron donor levels [113117], hybridizationof the O2p orbital with other orbitals (oxides [118,119], oxynitrides[120,121] and oxysuldes[18,122]), and the formation of solidsolutions [15,123,124]. However, there are still few reliable photo-catalysts for the one-step water splitting under visible light[123,124] due to the stringent requirements. An effective one-step water-splitting system that uses a solid solution photocata-lyst, i.e., (Ga1xZnx)(N1xOx) or (Zn1xGe)(N2Ox), [123,124] hasbeen described recently.

Water splitting under visible light illumination was demon-strated for the rst time in 2001 using a Z-scheme photocatalyticsystem consisting of SrTiO3 doped with Cr and Ta (denoted asSrTiO3:Cr/Ta) for H2 evolution, WO3 for O2 evolution, and aniodate/iodide (IO3/I) redox couple as electron mediator (Fig. 7)[11,39]. Prior to that, Z-scheme photocatalytic water splitting hadbeen demonstrated using a combination of Pt-loaded anatase TiO2and bare rutile TiO2 photocatalysts in the presence of an iodate/iodide (IO3/I) shuttle redox mediator [125]. Although thissystem operates only upon UV light irradiation (o400 nm) dueto the large band gap of the TiO2 photocatalysts, it eventuallyopened the way to achieve water splitting under visible light. Thekey was controlling the reactivity of electrons and holes with theredox mediator (IO3/I anions) and the water molecules, respec-tively. It is quite difcult to achieve the simultaneous evolution ofH2 and O2 in two-step water-splitting systems (Z-scheme) becausethe backward reactions of the redox mediator usually proceedreadily over both photocatalysts thus suppressing the forwardreactions (H2 and O2 evolutions) (Fig. 7). Highly efcient andselective O2 evolution was found to proceed over rutile TiO2 andPt-loaded WO3 (Pt/WO3) photocatalysts in an aqueous solution

containing IO3 anions as electron acceptors (Ox) according to thefollowing reactions (Eqs. (4)(6)):

h-eCB hVB (4)

IO33H2O6eCB-I6OH (5)

2H2O4hVB-O24H (6)

Z-scheme systems consisting of a modied BaZrO3BaTaO2Nsolid solution and a reversible donor/acceptor pair (i.e., IO3/I andFe3/Fe2) were constructed [126]. BaZrO3BaTaO2N having aband gap of 1.8 eV has sufcient potential to reduce and oxidizewater by absorbing visible photons of up to 660 nm (Fig.8). SrTiO3was selected as the mother structure and was doped with eithervanadium (and sodium) or rhodium to introduce visible lightsensitivity. By utilizing these two types of SrTiO3-based photo-catalysts, the simultaneous liberation of H2 and O2 in the presenceof (IO3)/I as a redox mediator under irradiation with only visiblelight (4420 nm) could be realized. QE values under irradiationwith 360 nm UV light and 420 and 480 nm visible light were 0.33,0.056, and 0.039%, respectively [127].

A fully integrated system of nanoscale photoelectrodesassembled from inorganic nanowires for direct solar water split-ting has been described [128]. The articial photosynthetic systemcomprises two semiconductor light absorbers with large surfacearea, an interfacial layer for charge transport, and spatiallyseparated cocatalysts to facilitate the water reduction and oxida-tion. Under simulated sunlight, a 0.12% solar-to-fuel conversionefciency is achieved, which is comparable to that of naturalphotosynthesis. A model Z-scheme system with two light-absorbing materials is chosen here to demonstrate the capabilityof an integrated nanostructure to use sunlight to split water [129].Earth abundant and stable semiconductors, silicon (Si) and TiO2,were chosen as the H2-generating photocathode and O2-generat-ing photoanode, respectively (Fig. 9). Upon illumination photo-excited electronhole pairs are generated in Si and TiO2, whichabsorb different regions of the solar spectrum (Fig. 9b), thephotogenerated electrons in the Si nanowires migrate to thesurface and reduce protons to generate H2; meanwhile the photo-generated holes in the TiO2 nanowires oxidize water to evolve O2.The holes from Si and electrons from TiO2 recombine at the ohmiccontact, completing the relay of the Z-scheme [129,130], similarto that of natural photosynthesis.

A visible light responsive plasmonic photocatalytic compositematerial has been designed by rationally selecting Au nanocrystalsand assembling them with TiO2-based photonic crystal substrates[131]. The design of the composite material was expected tosignicantly increase the Au surface plasmonic resonance (SPR)intensity and consequently boost the hot electron injection fromthe Au nanocrystals into the conduction band of TiO2, leading to aconsiderably enhanced water splitting performance of the

Fig. 7. Overview of water splitting on Z-scheme photocatalysis system with aniodate (IO3) and iodide (I) ion redox couple.Source: Ref. [103].

Fig. 8. Schematic illustration of spontaneous H2 and O2 evolution by Ru/Na,V-STOand Ru/Rh-STO under irradiation with visible light in the presence of IO3/I redoxmediator.Source: Ref. [127].


material under visible light. The results showed that under entiresolar light irradiation both, H2 and O2 were evolved at theexpected stoichiometric ratio with generation rates of27.9 mol h1 for H2 and 13.6 mol h1 for O2 within a 120 mintesting period.

2.2.4. Metal suldesCdS is probably the best studied metal sulde photocatalyst

due to its relatively narrow band gap (2.4 eV), it absorbs visiblelight at wavelengths of o510 nm [132]. The at-band potential ofCdS (0.87 V) [30] is sufciently high to reduce H2O, and the topof its valence band (1.5 V vs NHE) is theoretically suitable to allowthe oxidation of water. For 4 nm CdS nanoparticles, the lifetime ofphotogenerated charge carriers is on the order of 50 ps. However,prolonged irradiation of CdS suspensions leads to photocorrosionof CdS into Cd2 and S [31,133]. This reaction can be suppressedby the addition of reducing agents to the aqueous phase. Undervisible light irradiation, PtCdS powder evolved H2 from aqueousEDTA solutions with QE4%, but without Pt, the activity wasreduced by a factor of 10. Prolonged irradiation (44 h) leads to thedecomposition of the catalyst [134]. In 1984, Reber and co-workerspublished the most comprehensive study concerning photocata-lysis with CdSPt microcrystalline powders [135]. The best cata-lysts evolved H2 at 2.9 mmol h1 (QE25%) under irradiation withlight of 4300 nm in the presence of either S2 , SO32 , or S2/HPO2 as reducing agents. The activity dropped by 21% over thecourse of 6 days, likely because of the deactivitation of Pt due toformation of PtS species. Under solar irradiation, 50 mL/h of H2were produced with 0.4 g of catalyst suspension equal to anenergy efciency of 2%. In a follow-up study, the activity of CdScould be increased up to 357 mL/h1 (QE37%) of H2 in aqueousNa2S/Na2SO3 suspensions by doping CdS with 15 mol% ZnS and bythe addition of a Pt cocatalyst [136,137]. ZnS has a band gap of3.66 eV, which restricts light absorption to the UV (o340 nm).Similar to CdS, it undergoes photochemical decomposition into itscomponents when irradiated in the absence of sacricial electrondonors. The rst report on water splitting with ZnS was publishedby Yanagida in 1983 [138], who synthesized ZnS from ZnSO4 orZnCl2 and Na2S in water. With tetrahydrofuran or alcohols as thesacricial donor, this catalyst produced H2 gas under UV irradia-tion. The most comprehensive study on photocatalytic reactions ofZnS and ZnSPt was carried out by Reber et al. [133]. Under

irradiation with 4300 nm light and at 60 1C, ZnSPt catalyzed H2evolution with quantum yields of up to 90% from aqueoussolutions of sulde and sulte. To improve the visible lightabsorption of ZnS, Kudo's group has tested metal dopants (Cu,Ni, and Pb [139141]). Doping can move the absorption edge to500 nm in the case of Ni2 ions and to 550 nm in the case of Pb2

ions. Under visible light irradiation, a Ni2-doped catalyst pro-duced H2 from aqueous K2SO3/Na2S with QEs of up to 1.3% [140].Cu-doped ZnS gave QE3.7% under visible light irradiation for themolecular hydrogen formation from aqueous Na2SO3 [139]. Dop-ing of ZnS with variable amounts of AgInS2 or CuInS2 produces aseries of solid solutions that crystallize in the cubic zinc blende orhexagonal Wurtzite structure [142145]. Visible light irradiation ofthe respective Pt- or Ru-derivatized catalysts produced H2 witha QE of up to 7.5% from aqueous Na2S/Ka2SO3 [143]. For thePt-loaded Ag0.22In0.22Zn1.56S2 photocatalyst, a QE of 20% wasmeasured at 420 nm under solar irradiation conditions [145].Na14In17Cu3S35 xH2O evolved small quantities of H2 under visiblelight irradiation using Na2S as the sacricial electron donor, equalto QE3.7% at 420 nmwith SO32 , the QE dropped to 0.37%. Undervisible light irradiation, the Pt-modied catalyst showed goodphotocatalytic activity for H2 evolution from aqueous K2SO3solutions [146]. When supported on SiO2 and using uoresceinas a sensitizer, the catalyst produced H2 from aqueous EDTAsolutions under visible light illumination [147]. Bi2S3 has a bandgap of 1.28 eV; it produces H2 at intermediate rates from anaqueous sulde solution. Rates decline after 100 min, probablybecause of disulde formation. Platinization improves the activityby 25% [148].

2.2.5. GaN:ZnO solid solution photocatalystsDomen et al. demonstrated water splitting into H2 and O2

under visible light irradiation using a single (Ga1xZnx)(N1xOx)photocatalyst material with an average quantum yield of ca.0.14% [123]. Although (Ga1xZnx)(N1xOx) with a ZnO concen-tration above 75% can absorb long-wavelengths visible light[149,150], only GaN-rich (Ga1xZnx)(N1xOx) with a ZnO con-centration below 22% exhibited activity for overall water split-ting [151]. In the rst report, the origin of the visible lightabsorption was explained by band gap narrowing due to pdrepulsion between Zn 3d and N 2p electrons in the uppervalence band [123]. A solid solution of ZnO and Ge3N4

Fig. 9. Schematics of the asymmetric nanoscale tree-like heterostructures used for solar-driven water splitting. (a) Structural schematics of the nanotree heterostructure.The small diameter TiO2 nanowires (blue) were grown on the upper half of a Si nanowire (gray), and the two semiconductors absorb different regions of the solar spectrum.The two insets display the photoexcited electronhole pairs that are separated at the semiconductorelectrolyte interface to carry out water splitting with the help ofcocatalysts (yellow and gray dots on the surface). (b) Energy band diagram of the nanotree heterostructure for solar-driven water splitting. The photogenerated electrons inSi and holes in TiO2 move to the surface to perform water splitting, while the holes in Si and electrons in TiO2 recombine at the ohmic contact between the twosemiconductors. (For interpretation of the references to color in this gure legend, the reader is referred to the web version of this article.)Source: Ref. [130].


(Zn1xGe)(N2Ox) was also demonstrated to be an active photo-catalyst for the one-step overall water splitting under visiblelight irradiation [124].

2.2.5.1. Co-catalysts for efcient H2 production over GaN:ZnO. Mostsemiconductor photocatalysts that have been developed to dateneed to be loaded with suitable co-catalyst particles to promote H2production, whereas some photocatalysts with wide bandgaps candecompose water even without co-catalyst loading. To prevent thebackward reaction, transition metal oxides such as NiOx or RuO2have mainly been used as effective co-catalysts to decomposewater into H2 and O2. In NiOx co-catalysts, the Ni core functions asan electron pool that accepts photoexcited electrons from the bulkof the semiconductor and the NiO shell effectively suppresses thebackward reaction, while reduction of water to H2 can occur onthe surface [51]. However, the application of NiOx has been limitedto thermally stable materials such as metal oxides because loadingof NiOx co-catalysts requires relatively high temperatures foractivation. Therefore, NiOx cocatalysts appear to be unsuitable fornon-oxide photocatalysts. RhCr mixed oxide (Rh2yCryO3)nanoparticles as co-catalyst was investigated, which signicantlyenhances H2 production (Fig. 10) [14,15]. Dispersion of Rh2yCryO3nanoparticles on (Ga1xZnx)(N1xOx) produced an effectivecatalyst for H2 evolution exhibiting a quantum efciency of ca.2.5%, which was the highest quantum efciency reported forphotocatalytic water splitting under visible light irradiation [15].The Rh2yCryO3 revealed the important role of co-catalysts inachieving highly efcient water splitting over semiconductorphotocatalysts. Since the report of these Rh2yCryO3 co-catalysts,other researchers have reported various new co-catalyst materialsthat are effective for H2 evolution [152155].

Noble metal (core)/Cr2O3 (shell) nanoparticles were developedas a cocatalyst for water splitting. [156159] Noble metals such asRh and Pt were deposited on the surface of the photocatalyst byconventional photodeposition. The surfaces of the noble metalparticles were therefore coated with thin Cr2O3 layers, formingnoble metal (core)/Cr2O3 (shell) nanoparticles. Such core/shellnanoparticles can be readily prepared by the photoreduction ofCr(VI) ions using noble metal or metal oxide loaded photocatalystpowders in the absence of air [160]. Rh-loaded GaN:ZnO photo-catalysts exhibit little photocatalytic activity for water splittingdue to the rapid water formation from the evolved H2 and O2 onthe Rh nanoparticles. However, GaN:ZnO loaded with a Rh/Cr2O3core/shell co-catalyst particles yielded H2 and O2 evolution from

water. Electrochemical measurement on model electrodes with acore/shell structure revealed that the thin Cr2O3 layers do notinterfere with the proton reduction to H2 and that the protonreduction occurred at the Cr2O3/Pt interface [161]. On the otherhand, Cr2O3 layers effectively suppressed the reduction of O2 towater as illustrated in Fig. 11.

The co-loading of Rh (core)/Cr2O3 (shell) and Mn3O4 nanopar-ticles as H2 and O2 evolution promoters, respectively, on GaN:ZnOresulted in higher photocatalytic activities as compared withMn3O4 nanoparticles modied with either Rh/Cr2O3 or Mn3O4[162]. A gradual size control of poly(N-vinyl-2-pyrrolidone)-pro-tected Rh nanoparticles was applied for the use as cocatalysts onthe photocatalyst (Ga1xZnx)(N1xOx) achieving overall watersplitting under visible light illumination. Moreover, it was illu-strated that smaller Rh cores yielded higher activities than largerones. These results indicate that Rh loaded on GaN:ZnO, additionalmodied with Cr2O3, achieve the functionality as cocatalysts topromote the water splitting by GaN:ZnO [163].

Fig. 10. Photocatalytic performance of the photocatalyst (Ga1xZnx)(N1xOx) loaded with mixed oxides of rhodium and chromium in overall water splitting.Source: Ref.[15].

Fig. 11. Schematic illustration of H2 evolution on core/shell-structured nanoparti-cles (with a noble metal or metal oxide core and a Cr2O3 shell) as a cocatalyst forphotocatalytic overall water splitting.Source: Ref. [160].


2.2.5.2. Mechanism of photocatalytic water splitting. Many studieshave focused on the development of materials that are suitable forthe visible-light-driven overall water splitting by addressing lightabsorption properties, band edge position, crystallographic quality,particle morphology, and phase purity [160]. However, it is difcultto understand what factor(s) dominate(s) the net photocatalyticactivity based on the above physical properties because thephotocatalytic reactions proceed through a complicated sequenceof competing multistep processes. Aspects of the photocatalyticwater splitting mechanism on GaN:ZnO powder modied with aRhCr mixed oxide cocatalyst were investigated with respect to theeffects of cocatalyst loading, light intensity, hydrogen/deuteriumisotopes, and reaction temperature on the photocatalytic activity[162]. The water splitting rate with the optimally modiedphotocatalyst was found to be proportional to the light intensityunder solar-equivalent or weaker irradiation, indicating that theaccumulation of photoexcited electrons and holes was negligible.The HD isotope effect on the overall water splitting was signi-cantly lower than previously reported values for photocatalytic andelectrochemical H2 evolution reactions. The apparent activationenergy for the overall water splitting was as low as 8 kJ mol1 andwas unchanged by the addition of electron donors or acceptors [160].These results reect a shortage of photoexcited carriers available forsurface redox reactions under steady light irradiation. In summary,the experimental results indicate that the balance between the ratesof redox reactions on the photocatalyst surface and the carriergeneration/recombination in the photocatalyst bulk determines thesteady state charge concentration in the photocatalyst, that is,developing both, the photocatalyst and the cocatalyst is important.The proposed kinetic model of photocatalytic water splitting also

suggests that the reaction probability of photoexcited holes for O2evolution versus their recombination with intrinsic electrons in thephotocatalyst determines the water splitting activity of GaN:ZnO. Itwould be natural to expect that loading both, H2 and O2 evolutioncocatalysts, onto the same photocatalyst would improve the watersplitting activity, compared to that for photocatalysts modied witheither an H2 or an O2 evolution cocatalyst.

It is easy to imagine that the two different cocatalysts wouldseparately facilitate H2 and O2 evolution, thereby promoting theoverall water splitting in harmony. Recently, GaN:ZnO loaded withRh/Cr2O3 (core/shell) and Mn3O4 nanoparticles was demonstratedas H2 and O2 evolution promoters, respectively, under visible lightirradiation [164]. The goal of the respective study was to generateboth, H2 and O2 evolution sites, separately on the same photo-catalyst surface. The preparation method developed by Domen'sgroup is a stepwise deposition involving the adsorption of MnOnanoparticles followed by calcination to obtain crystallized Mn3O4nanoparticles and a subsequent photodeposition of Rh/Cr2O3(core/shell) nanoparticles, as shown schematically in Fig. 12a.As mentioned earlier, core/shell structured Rh/Cr2O3 nanoparticlesprovide active sites for H2 evolution [164]. On the other hand,photoelectrochemical analysis revealed that Mn3O4 nanoparticleson GaN:ZnO promote the photooxidation of water. Finally, overallwater splitting was attempted using the as-prepared samplesunder visible light irradiation. As expected, the activity of GaN:ZnO modied with both Rh/Cr2O3 and Mn3O4 provided a higheractivity than modication with either Rh/Cr2O3 or Mn3O4, asshown in Fig. 12b. The proposed reaction scheme of overall watersplitting on GaN:ZnO modied with Rh/Cr2O3 and Mn3O4 isdepicted in Fig. 12c [160].

Fig. 12. (a) Scheme for the preparation of GaN:ZnO loaded with both Mn3O4 and core/shell-structured Rh/Cr2O3. (b) Photocatalytic activity of GaN:ZnO modied withdifferent cocatalysts under visible light (4420 nm). (c) Illustration of the reaction scheme for overall water splitting on GaN:ZnO modied with Mn3O4 and core/shell-structured Rh/Cr2O3.Source: Ref. [160].


2.2.6. Overall water splitting employing dye-sensitized solar cells toharvest visible light

The basic principle of the dye-sensitized photocatalytic H2production from water is shown in Fig. 13. Photoexcitation of thedye adsorbed onto a semiconductor leads to the injection ofelectrons into the conduction band of this semiconductor. Theelectrons are consumed by the reduction of water to produce H2.The oxidized dye molecules are subsequently reduced and thusregenerated by accepting electrons from a suitable electron donor[165,166]. Photocatalytic H2 production systems in which ruthe-nium(II) complex dyes sensitize wide band gap semiconductors tovisible light have been the focus of intensive research for manyyears. In the early 1980s, Gratzel and co-workers [167,168]succeeded in decomposing water by visible light using tris(2,2'-bipyridine)ruthenium(2)ion (Ru(bpy)32) and its amphiphilicderivatives as sensitizers. Pt/RuO2-loaded TiO2 particles provedparticularly effective in these systems, acting as electron mediatorsand catalysts for H2 formation process. Nakahira et al. found thatPt/TiO2 sensitized with a polymer-pendant Ru(bpy)32 complexwas effective in H2 evolution in the presence of the sacricialdonor (EDTA) under visible-light irradiation [169]. Hirano et al.found that, when tris(bipyrimidine)ruthenium(II) (Ru(bpym)32)was used as the visible-light sensitizer, Pt/TiO2 showed muchhigher efciency for H2 production than with Ru(bpy)32 [170].Pt/TiO2/p-Ru(bpy)32 anchored through phosphonate groups exhi-bited higher photocatalytic activity for H2 production from waterthan Pt/TiO2/c-Ru(bpy)32 anchored through carboxylate groups[171]. Because of the more rapid regeneration of p-Ru(bpy)32 ascompared with c-Ru(bpy)32 , the phosphonate group seemed to bebetter than the carboxylate group as a ruthenium sensitizerlinkage to the TiO2 surface in an aqueous environment. The effectof the energy gap between the I3/I redox potential and thehighest occupied molecular orbital (HOMO) level of the dyes onthe photocatalytic activity of Ru complex dye-sensitized Pt/TiO2 inwater/acetonitrile solutions was studied [172]. The inuence ofdifferent ruthenium(II)bipyridyl complexes on the photocatalyticH2 evolution rate on TiO2 under visible light illumination wasinvestigated [173,174]. When compared to Ru(bpy)2(him)2-NO3-and Ru(dcbpy)2(NCS)2-sensitized Pt/TiO2, Ru2(bpy)4L1-PF6-sensitized Pt/TiO2 displayed higher photocatalytic efciency forH2 evolution. This may be related to the dynamic equilibriumbetween the linkage of the ground-state dye to TiO2 and theseparation of the oxidized dye from TiO2. The addition of Al2O3overlayer on Ru(bpy)32-sensitized TiO2 was signicantly increasedthe visible-light-sensitized activity for H2 production [175]. Thisresulted in an enhanced photocatalytic H2 production rate uponvisible light irradiation [176]. Reisner et al. constructed a specialsystem consisting of [NiFeSe]-hydrogenase attached to Ru dye-sensitized TiO2, with triethanolamine (TEA) being the sacricial

electron donor [177]. This system showed a high and stablephotocatalytic activity for H2 generation under visible light illu-mination. Also, the photocatalytic efciency of SnO2 could begreatly improved using Ru(bpy)32 sensitization [178]. The highestphotocatalytic activity for H2 evolution, with a quantum yield of2.40% upon illumination was observed for a Pt/SnO2/RuO2Ru(bpy)32MV2EDTA system. The dynamics of photoexcited Ru(bpy)32 intercalated into K4Nb6O17 interlayers was investigated[179]. Because of the fast and efcient electron transfer betweenRu(bpy)32 and K4Nb6O17, the transient bleaching of the Ru(bpy)32

band showed a fast and non-exponential decay. This differedfrom the behavior of Ru(bpy)32 in water, and thus explained thephotocatalytic H2 evolution over Ru(bpy)32 intercalated K4Nb6O17from aqueous solutions containing appropriate electron donors[180]. Moreover, different layered oxide semiconductors werestudied as building blocks for the visible-light induced H2 produc-tion from water using ruthenium complexes as photosensitizers[164,181185]. Ru(bpy)32-sensitized K4Nb6O17 nanoscrolls werefound to exhibit higher photocatalytic activity than Ru(bpy)32-sensitized lamellar K4Nb6O17 and TiO2. This was primarilyexplained by the high surface area of the nanoscrolls and by theirexcellent ability to bind Ru(bpy)32 thus leading to the faciletransfer of electrons from the sensitizer to the Pt catalyst islandsvia the single-crystalline nanoscrolls. The authors claimed that theplatinized K4Nb6O17 nanoscrolls were slightly better electrontransfer mediators than acid-restacked HCa2Nb3O10 nanosheets.The apparent quantum yield of visible-light photocatalytic H2production over Pt/H4Nb6O17 nanoscrolls was reported to be 25%upon 450 nm excitation, when sensitized by Ru(bpy)32 complexes[182,183]. The simplest system for exploring the idea of visiblelight water splitting with a sensitized oxide semiconductor is aphotoelectrochemical cell [181] thus avoiding possible complica-tions through the undesired H2O2 recombination by generating

Fig. 13. Basic principle of dye-sensitized photocatalytic H2 production from water.Source: Ref. [165].

Fig. 14. (top) Schematic diagram of a water-splitting dye sensitized solar cell. Theinset illustrates a sensitizer-capped IrO2 nH2O catalyst particle in the mesopores ofthe TiO2 electrode lm. (bottom left) Current transient obtained upon visible lightillumination. (bottom right) Energy level diagram showing the rates of forward andback electron transfer from and to the sensitizer molecule.Source: Ref. [181].


H2 in a physically separated cathode compartment. A scheme ofsuch a cell is shown in Fig. 14with the photoanode being amesoporous TiO2 electrode [186], like in a dye-sensitized solarcell [187]. The dye is a monolayer of sensitizer capped IrO2 nH2Oparticles. The [Ru(bpy)3]2 sensitizer is modied with bothphosphonate and malonate ligands in the 4-positions of the 2,20-bipyridyl ligands in order to adsorb strongly to TiO2 and onIrO2 nH2O, respectively. The cathode is a Pt wire electrode.Because the potential of electrons in trap states below the anataseconduction band is not sufciently negative to reduce water, a biasvoltage (330 mV) must be applied for water splitting to occur inthis cell. In a conventional DSSC, the anode and cathode areconnected by a I/I3 redox couple, which very rapidly regeneratesthe Ru(II) state of the sensitizer following electron injection intothe TiO2 anode lm. The observed low quantum efciencies can beunderstood in terms of three problems that can, in principle, beaddressed by an improved design at the molecular level [181].The simplest of these problems is to make catalyst particles thatare connected to only one sensitizer molecule, so that eachsensitizer can bind both TiO2 and IrO2 nH2O. A second problemis to slow down the back electron transfer reaction by changingthe distance between redox partners, and a third is to speed up theelectron transfer from Ir(IV) to Ru(III). The photocurrent transients

in this system (Fig. 14) show an initial spike that decays over atimescale of seconds, indicating a certain degree of polarization ofthe mesoporous TiO2 electrode. Because the O2 evolution reactiongenerates protons, the current decay may result from a local pHdrop in the TiO2 lm, which decreases the overpotential for wateroxidation. A simple solution to this problemwould be to nd moreeffective buffers that can penetrate into the porous lm.

Recently, a molecular water oxidation (Fig.15) Ru catalyst (1)(silane group, Ru(II)(bda)(4-picoline) L (H2bdabipyridine-dicar-boxylic acid; LN-(3-(triethoxysilyl) propyl) isonicotinamide)(2))has been synthesized and immobilized together with a molecularphotosensitizer [Ru(bpy)2(4,4-(PO3H2)2bpy)]Br2 (2) on nanostruc-tured TiO2 particles xed on conducting FTO glass, forming aphotoactive anode (TiO2 (12))[184]. By using TiO2 (12) as theworking electrode in a three-electrode photoelectrochemical cell(PEC), visible light driven water splitting has been successfullydemonstrated in an aqueous phosphate buffer solution, with O2and H2 bubbles evolved from the working and the counterelectrode, respectively. A high photocurrent density of more than1.7 mA cm2 has been achieved in this set-up.

2.2.7. Electron mediators for overall water splittingWater splitting using photocatalysts has a serious problem in that

the obtained gas will be a mixture of H2 and O2. However, it would bepossible to generate H2 separately from O2 using electron mediators.It is therefore important to nd new electron mediators for theimprovement of the efciency of Z-scheme systems as well as forthe development of new systems. Kudo's group has reported a novelZ-scheme water splitting system without an electron mediator [188].Fujihara et al. have succeeded in the overall water splitting under UVlight irradiation by the Z scheme photocatalyst system using twocompartments connected by a Pt wire with bromide and iron ionsacting as electron mediators [189]. The separation of molecular H2 andO2 is achieved under visible light irradiation by O2 evolution on WO3from a ltered aqueous solution containing iron ions following H2evolution on Pt/SrTiO3:Rh [190]. Many transition metal complexeshave been reported as electron mediators in dye-sensitized solar cells[191193]. Gratzel et al. have found that the Co-complexes withimine-group ligands are comparable electron mediators to the well-known I3/I redox couple in dye-sensitized solar cells or on inorganicphotocatalysts [53,191,193]. In addition, Elliot et al. have reported Co-complexes to be effective mediators when combined with otherligands [192]. There are some analogies between an electron mediatorof the Z-scheme photocatalyst system and that of a dye-sensitizedsolar cell, the required redox potential, the reversibility, and theelectron transfer between the solid and the liquid phase (Fig.16). [Co(bpy)3]3 /2 and [Co(phen)3]3 /2 redox couples were reported toact as electron mediators for the overall water splitting under

Fig. 15. Schematic illustration of a molecular device employing a photoanode co-adsorbed with a photosensitizer [Ru(bpy)2(4,4-(PO3H2)2bpy)]Br2 (1) and a mole-cular Ru catalyst (2) on nanostructured TiO2 (TiO2(12)), and a passive Pt cathode,for visible light driven water splitting in aqueous solution.Source: Ref. [184].

Fig. 16. Overall water splitting using a Z-scheme photocatalyst system.Source: Ref. [194].


visible light irradiation in Z-scheme photocatalyst systems com-posed of Ru/SrTiO3:Rh and BiVO4 powders. [194]

3. Main challenges and opportunities

Signicant progress has beenmade in recent years for the develop-ment of novel nano- and photocatalysts [3,195199]. Nevertheless, theefciency of photocatalysis under visible light, must be improved inorder to meet engineering requirements. Furthermore, the stabilityand cost of these materials should also be carefully considered. It isthus a challenge of great importance to identify and design newsemiconductor materials that are efcient, stable, and abundant [3].The rst property relevant to the photocatalytic activity of a semi-conductor is its energy band conguration, which determines theabsorption of incident photons, the photoexcitation of electron-holepairs, the migration of the charge carriers, and the redox capabilitiesof excited-state electrons and holes [27]. Therefore, energy bandengineering is a fundamental aspect for the design and fabricationof semiconductor photocatalysts. Regarding optical absorption, directand narrow band gap semiconductors are more likely to exhibit highabsorbance and is suitable for the efcient harvesting of low energyphotons. However, the recombination probability for photo-excitedelectron-hole pairs is rather high in these types of semiconductors,and the band-edge positions are frequently incompatible with theelectrochemical potential that is necessary to trigger specic redoxreactions [3]. Some renowned photocatalysts such as TiO2 and WO3are in fact indirect band gap semiconductors, which, however, exhibita rather abrupt absorption onset similar to direct band gap semi-conductors. In order to modulate the band gap and band-edgepositions in a precise manner, both, doping with different elementsand solid-solution strategies have been extensively investigated andapplied [200204]. Furthermore, the inter-particle electronic couplingof semiconductor nanocrystals has been found to be effective forenergy band reconstruction and for band gap narrowing [205]. Theutilization efciency of incident photons in photocatalytic semicon-ductor materials can be improved by methods other than energy bandengineering, such as the improvement of light sensitization by theinclusion of quantum dots [206], the plasmonexciton couplingbetween anchored noble metal nanoparticle co-catalysts and the hostsemiconductor [207], and the photon coupling in photonic semicon-ductor crystals [208]. Another key issue inuencing the photocatalyticcapability of a semiconductor is the nature of its surface/interfacialchemistry. The surface energy as well as the chemisorption propertiesplay crucial roles in the transfer of electrons and energy betweensubstances at the interface, in governing the selectivity, the rate andthe overpotential of redox reactions on the photocatalyst surface, andin determining the susceptibility of the photocatalyst toward photo-corrosion [209]. In general, a higher surface energy should yield highercatalytic activity. Recently, much interest has been focused on researchinto semiconductor crystals with morphologies providing large per-centages of highly reactive facets [210]. Attempts to deliberatelyfabricate such materials are usually challenged by the thermodynamicgrowth mechanisms of the crystals. In this way, it is possible to growspecic nanostructured materials with surfaces comprising nearly100% high-energy facets with the examples including ultra-thin sheetsand highly symmetric polyhedral particles.

During the search for visible-light responsive photocatalysts,signicant efforts have been devoted to the development of activesites on the photocatalysts and to the elucidation of the underlyingreaction mechanisms, leading to signicant progress in the eld ofheterogeneous photocatalysis for water splitting, especially duringthe last 5 years [15,153161]. Several promising systems, includingRh2yCryO3-loaded GaN:ZnO (a one-step water splitting system)and a two-step system consisting of Pt/ZrO2/TaON and Pt/WO3have been developed employing IO3/I shuttle redox mediators,

with respective apparent quantum yields of about 5.1% uponillumination at 410 nm and 6.3% upon illumination at 420 nm,respectively. However, research continues to pursue more activephotocatalytic systems capable of harvesting more visible photons.As shown in Fig. 17, the solar energy conversion efciencyincreases once the overall water splitting can be achieved uponlonger wavelengths irradiation. Needless to say, this is because thenumber of available photons in the solar spectrum increases withincreasing wavelength. To provide one-third of the projectedenergy needs of the human society in 2050 from solar energy,preliminary estimation by Doman et al. suggest that approxi-mately 10,000 solar plants need to be built with each plant being5 km5 km in area and a solar energy conversion efciency of10%. The total required area for these constructions, 250,000 km2,corresponds to 1% of the earth's desert area; 570 t of H2 gas wouldbe produced per day, assuming an integrated solar energy of AM1.5G irradiation for a day with correction for sunlight angle. ThisH2 would be available for use as a recyclable reactant in fuel cellsand as a raw material for the production of important chemicalssuch as methanol and so on. Of course, a technology to separatesimultaneously produced H2 and O2 will also be required.

4. Conclusions and outlook

Over 140 metal oxides, Perovskites and oxynitrides are known tocatalyze the photochemical water-splitting reaction. Even though theprinciple activity controlling factors in the employed semiconductor-heterostructures have been identied, many aspects of the function ofinorganic photocatalysts are still unclear. Metal compounds with d0

ions (Ti, Zr, Nb, and Ta) and d10 ions (Ga, In, Ge, Sn, and Sb) bothexhibit activities for the overall photochemical splitting of water.Oxides are dominant, but nitrides and oxynitrides have also beenshown to catalyze the reaction. Most importantly, the molecularmechanism of water reduction and oxidation on the semiconductorsurface has not yet been elucidated in sufcient details. Manyquestions concerning the charge transfer between semiconductorsand cocatalysts, and its dependence on the structural and electronicfeatures of the interface still remain unanswered. The effects ofvariable material preparations and surface impurities on the catalyticactivity of semiconductors have not been fully considered. These R&Dareas represent signicant opportunities for improving water splittingphotocatalysts. The development of better photocatalysts is also going

Fig. 17. Calculated solar energy conversion efciency as a function of wavelengthfor the overall water splitting using photocatalysts with various quantum efcien-cies. The solar irradiance used for these calculations is taken from AM 1.5G data.Source: Ref. [160].


to benet from the recent progress in nanoscience. Quantum sizeeffects can now be used to tailor both, the electronic structure and thereactivity of nanostructures while synthetic methods can be employedfor controlling the morphology of catalysts down to the nanoscale tofurther raise the efciency of photochemical water splitting systems.Further efforts concerning the development of novel co-catalysts mayfurthermore provide additional breakthroughs in obtaining highlyefcient photocatalysts. Such progress in the research is currentlyunder way along with new synthetic directions for the photocatalystpreparation and novel system development processes.

Acknowledgment

A.A. Ismail acknowledges the Alexander von Humboldt Foundation(AvH) for granting him a research fellowship (Renewed Research Stay).D.W. Bahnemann acknowledges a megagrant from the RussianMinistry of Science supporting the work for this publication.

References

[1] A. Fujishima, K. Honda, Electrochemical photolysis of water at a semicon-ductor electrode, Nature 238 (1972) 37.

[2] A. Fujishima, K. Honda, Electrochemical evidence for the mechanism of theprimary stage of photosynthesis Bull, Chem. Soc. Jpn 44 (1971) 11481150.

[3] H. Tong, S. Ouyang, Y. Bi, N. Umezawa, M. Oshikiri, J. Ye, Nano-photocatalyticmaterials: possibilities and challenges, Adv. Mater. 24 (2012) 229251.

[4] T. Ikeda, A. Xiong, T. Yoshinaga, K. Maeda, K. Domen, T. Teranishi, Polyolsynthesis of size-controlled Rh nanoparticles and their application tophotocatalytic overall water splitting under visible light, J. Phys. Chem.C117 (2013) 24672473.

[5] J. Mndez-Ramos, J.C. Ruiz-Morales, P. Acosta-Moraa, J. del-Castilloc,A.C. Yanes, Rare-earth doped nano-glass-ceramics for extending spectral responseof water-splitting semiconductor electrodes by high intense UV-blue up-conver-sion: turning the sun into blue, J. Power Sources 238 (2013) 313317.

[6] M. Law, L.E. Greene, J.C. Johnson, R. Saykally, P.D. Yang, Nanowire dye-sensitized solar cells, Nat. Mater. 4 (2005) 455459.

[7] X. Chen, S. Shen, L.S. Guo, S. Mao, Semiconductor-based photocatalytichydrogen generation, Chem. Rev. 110 (2010) 65036570.

[8] Y. Moriya, T. Takatab, K. Domen, Recent progress in the development of (oxy)nitride photocatalysts for water splitting under visible-light irradiation,Coord. Chem. Rev. 257 (2013) 19571969.

[9] A.J. Bard, M.A. Fox, Articial photosynthesis: solar splitting of water tohydrogen and oxygen, Acc. Chem. Res. 28 (1995) 141145.

[10] K. Maeda, D. Lu, K. Domen, Direct water splitting into H2 and O2 was recentlyachieved by modied TaON photocataylst under visible light irradiation,Chem. Eur. J. 19 (2013) 49864991.

[11] H. Kato, K. Asakura, A. Kudo, Highly efcient water splitting into H2 and O2over Lanthanum-doped NaTaO3 photocatalysts with high crystallinity andsurface nanostructure, J. Am. Chem. Soc. 125 (2003) 30823089.

[12] J.F. Reber, K. Meier, Photochemical production of hydrogen with zinc suldesuspensions, J. Phys. Chem. 88 (1984) 59035913.

[13] K. Maeda, K. Domen, New non-oxide photocatalysts designed for overallwater splitting under visible light, J. Phys. Chem. C 111 (2007) 78517861.

[14] K. Maeda, K. Teramura, D.L. Lu, T. Takata, N. Saito, Y. Inoue, K. Domen,Characterization of RhCr mixed-oxide nanoparticles dispersed on(Ga1xZnx)(N1xOx) as a cocatalyst for visible-driven overall water splitting,J. Phys. Chem. B 110 (2006) 1375313758.

[15] K. Maeda, K. Teramura, D.L. Lu, T. Takata, N. Saito, Y. Inoue, K. Domen,Photocatalyst releasing hydrogen from water, Nature 440 (2006) 295.

[16] J.Y. Ying, C.P. Mehnert, M.S. Wong, Synthesis and applications ofsupramolecular-templated mesoporous materials, Angew. Chem. Int. Ed. 38(1999) 5677.

[17] R. Mokaya, Ultrastable mesoporous aluminosilicates by grafting routes,Angew. Chem. Int. Ed. 38 (1999) 29302934.

[18] A. Ishikawa, T. Takata, T. Matsumura, J.N. Kondo, M. Hara, H. Kobayashi,K. Domen, Oxysuldes Ln2Ti2S2O5 as stable photocatalysts for water oxida-tion and reduction under visible-light irradiation, J. Phys. Chem. B 108 (2004)26372642.

[19] R. Konta, T. Ishii, H. Kato, A. Kudo, Photocatalytic activities of noble metal iondoped SrTiO3 under visible light irradiation, J. Phys. Chem. B 108 (2004)89928995.

[20] Z.C. Liu, W.H. Shen, W.B. Bu, H.R. Chen, Z.L. Hua, Low-temperature formationof nanocrystalline -SiC with high surface area and mesoporosity viareaction of mesoporous carbon and silicon powder, Microporous Mesopor-ous Mater. 82 (2005) 137145.

[21] N.S. Lewis, D.G. Nocera, Powering the planet: chemical challenges in solarenergy utilization, Proc. Natl. Acad. Sci. U.S.A 103 (2006) 1572915735.

[22] N.S. Lewis, G. Crabtree, A.J. Nozik, M.R. Wasielewski, A.P. Alivisatos, BasicResearch Needs for Solar Energy Utilization, U.S. Department of Energy,Washington, DC, 2005.

[23] W.U. Huynh, J.J. Dittmer, A.P. Alivisatos, Hybrid nanorod-polymer solar cells,Science 295 (2002) 24252427.

[24] M. Gratzel, Photoelectrochemical cells, Nature 414 (2001) 338344.[25] M.A. Green, K. Emery, D.L. King, S. Igari, W. Warta, Solar cell efciency

tables (version 26), Prog. Photovolt: Res. Appl. 13 (387392) (2005).[26] A.J. Frank, N. Kopidakis, J. van de Lagemaat, Electrons in nanostructured TiO2

solar cells: transport, recombination and photovoltaic properties, Coord.Chem. Rev. 248 (2004) 11651179.

[27] D.E. Scaife, Oxide semiconductors in photoelectrochemical conversion ofsolar energy, Solar Energy 25 (1980) 41.

[28] Z.G. Zou, H. Arakawa, Direct water splitting into H2 and O2 under visible lightirradiation with a new series of mixed oxide semiconductor photocatalysts, J,Photochem. Photobiol. A 158 (2003) 145162.

[29] Z.G. Zou, J.H. Ye, K. Sayama, H. Arakawa, Direct splitting of water undervisible light irradiation with an oxide semiconductor photocatalyst, Nature414 (2001) 625627.

[30] D. Meissner, R. Memming, B. Kastening, Photoelectrochemistry of cadmiumsulde. 1. Reanalysis of photocorrosion and at-band potential, J. Phys. Chem.92 (1988) 34763483.

[31] D. Meissner, R. Memming, B. Kastening, D. Bahnemann, Fundamental problems ofwater splitting at cadmium sulde, Chem. Phys. Lett. 127 (1986) 419423.

[32] A.J. Nozik, R. Memming, Physical chemistry of semiconductor liquid inter-faces, J. Phys. Chem. 100 (1996) 1306113078.

[33] Y.Y. Gurevich, Y.V. Pleskov, Semiconductor Photoelectrochemistry, Consul-tants Bureau, New York (1986) 422.

[34] M.A. Butler, D.S. Ginley, Prediction of atband potentials at semiconductorelectrolyte interfaces from atomic electronegativities, J. Electrochem. Soc.125 (1978) 228232.

[35] A.L. Linsebigler, G. Lu, J.T. Yates, Photocatalysis on TiO2 surfaces: principles,mechanisms, and selected results, Chem. Rev. 95 (1995) 735758.

[36] M. Kaneko, I. Okura, Photocatalysis: science and technology, Springer-Verlag,New York, 2002.

[37] N. Serpone, G. Sauve, R. Koch, H. Tahiri, P. Pichat, P. Piccinini, E. Pelizzetti,H. Hidaka, Standardization protocol of process efciencies and activationparameters in heterogeneous photocatalysis: relative photonic efcienciesr, J. Photochem. Photobiol., A 94 (1996) 191203.

[38] K. Rajeshwar, N.R. de Tacconi, C.R. Chenthamarakshan, Semiconductor-basedcomposite materials: preparation, properties and performance, Chem. Mater.13 (2001) 27652782.

[39] R. Abe, K. Sayama, H. Sugihara, Development of new photocatalytic watersplitting into H2 and O2 using two different semiconductor photocatalysts anda shuttle redox mediator IO3/I , J. Phys. Chem. B 109 (2005) 1605216061.

[40] S. Sakthivel, H. Kisch, Daylight photocatalysis by carbon-modied titaniumdioxide, Angew. Chem., Int. Ed. 42 (2003) 49084911.

[41] R. Asahi, T. Morikawa, T. Ohwaki, K. Aoki, Y. Taga, Visible-light photocatalysisin nitrogen-doped titanium oxides, Science 293 (2001) 269271.

[42] S.-H. Liu, H.-R. Syu, One-step fabrication of N-doped mesoporous TiO2nanoparticles by self-assembly for photocatalytic water splitting undervisible light, Appl. Energy 100 (2012) 148154.

[43] T. Ohno, T. Mitsui, M. Matsumura, Photocatalytic activity of S-doped TiO2photocatalyst under visible light, Chem. Lett. 32 (2003) 364365.

[44] H.G. Kim, D.W. Hwang, J.S. Lee, A.n. Undoped, Single-phase oxide photocata-lyst working under visible light, J. Am. Chem. Soc. 126 (2004) 89128913.

[45] T.A. Kandiel, A. Feldhoff, L. Robben, R. Dillert, D.W. Bahnemann, Tailoredtitanium dioxide nanomaterials: anatase nanoparticles and brookite nanor-ods as highly active photocatalysts, Chem. Mater. 22 (2010) 20502060.

[46] A.A. Ismail, T.A. Kandiel, D.W. Bahnemann, Novel (and better?) titania-basedphotocatalysts: brookite nanorods and mesoporous structures, J. Photochem.Photobiol. A 216 (2010) 183193.

[47] T.A. Kandiel, A.A. Ismail, D.W. Bahnemann, Mesoporous TiO2 nanostructures:a route to minimize Pt loading on titania photocatalysts for hydrogenproduction, Phys. Chem. Chem. Phys. 12 (2011) 2015520161.

[48] X.-Y. Zhang, H.-P. Li, X.-L. Cui, Y. Lin, Graphene/TiO2 nanocomposites:synthesis, characterization and application in hydrogen evolution fromwater photocatalytic splitting, J. Mater. Chem. 20 (2010) 28012806.

[49] M.S. Wrighton, A.B. Ellis, P.T. Wolczanski, D.L. Morse, H.B. Abrahamson,D.S. Ginley, Strontium titanate photoelectrodes. Efcient photoassisted electro-lysis of water at zero applied potential, J. Am. Chem. Soc. 98 (1976) 27742779.

[50] K. Ohashi, J. McCann, J.O.M. Bockris, Stable photoelectrochemical cells for thesplitting of water, Nature 266 (1977) 610611.

[51] K. Domen, S. Naito, M. Soma, T. Onishi, K. Tamaru, Photocatalytic decom-position of water vapour on an NiOSrTiO3 catalyst, J. Chem. Soc.,Chem.Commun. 12 (1980) 543544.

[52] K. Domen, S. Naito, T. Onishi, K. Tamaru, Photocatalytic decomposition ofliquid water on a NiO-SrTiO3 catalyst, Chem. Phys. Lett. 92 (1982) 433434.

[53] K. Sayama, K. Mukasa, R. Abe, Y. Abe, H. Arakawa, Stoichiometric watersplitting into H2 and O2 using a mixture of two different photocatalysts andan IO3/I shuttle redox mediator under visible light irradiation, Chem.Commun. (2001) 24162417.

[54] A.A. Ismail, Single-step synthesis of a highly active photocatalyst for oxida-tion of trichloroethylene, Appl. Catal. B: Environ. 85 (2008) 3339.


[55] A.A. Ismail, L. Robben, D.W. Bahnemann, Study of the efciency of UV andvisible-Light photocatalytic oxidation of methanol on mesoporous RuO2TiO2 nanocomposites, ChemPhysChem 12 (2011) 982991.

[56] J. Kim, D.W. Hwang, H.G. Kim, S.W. Bae, J.S. Lee, W. Li, S.H. Oh, Highlyefcient overall water splitting through optimization of preparation andoperation conditions of layered perovskite photocatalysts, Top Catal. 35(2005) 295303.

[57] H.G. Kim, D.W. Hwang, J. Kim, Y.G. Kim, J.S. Lee, Highly donor-doped (110)layered perovskite materials as novel photocatalysts for overall watersplitting, Chem. Commun. (1999) )10771078.

[58] R. Abe, M. Higashi, K. Sayama, Y. Abe, H. Sugihara, Photocatalytic activity ofR3MO7 and R2Ti2O7 (RY, Gd, La; MNb, Ta) for water splitting into H2 andO2, J. Phys. Chem. B 110 (2006) 22192226.

[59] H. Kato, K. Asakura, A. Kudo, Highly efcient water splitting into H2 and O2over lanthanum-doped NaTaO3 photocatalysts with high crystallinity andsurface nanostructure, J. Am. Chem. Soc. 125 (2003) 30823089.

[60] D.W. Hwang, H.G. Kirn, J.S. Lee, J. Kim, W. Li, S.H. Oh, Photocatalytic hydrogenproduction fromwater over M-doped La2Ti2O7(MCr, Fe) under visible lightirradiation (4420 nm), J. Phys. Chem. B 109 (2005) 20932102.

[61] X. Chen, T. Yu, X. Fan, H. Zhang, Z. Li, J. Ye, Z. Zou, Enhanced activity ofmesoporous Nb2O5 for photocatalytic hydrogen production, Appl. Surf. Sci.253 (2007) 85008506.

[62] H.Y. Lin, H.C. Huang, W.L. Wang, Preparation of mesoporous InNb mixedoxides and its application in photocatalytic water splitting for hydrogenproduction, Microporous Mesoporous Mater. 115 (2008) 568575.

[63] K. Domen, A. Kudo, M. Shibata, A. Tanaka, K. Maruya, T. Onishi, Novelphotocatalysts, ion-exchanged K4Nb6O17, with a layer structure, J. Chem.Soc., Chem. Commun. 23 (1986) 17061707.

[64] L. Yan, J. Zhang, X. Zhou, X. Wu, J. Lan, Y. Wang, G. Liu, J. Yu, L. Zhi, Crystallinephase-dependent photocatalytic water splitting for hydrogen generation onKNbO3 submicro-crystals, Int. J. Hydrog. Energy 38 (2013) 35543561.

[65] D. Chen, J. Ye, Selective-synthesis of high-performance single-crystallineSr2Nb2O7 nanoribbon and SrNb2O6 nanorod photocatalysts, Chem. Mater. 21(2009) 23272333.

[66] H.G. Kim, D.W. Hwang, J. Kim, Y.G. Kim, J.S. Lee, Highly donor-doped (110)layered perovskite materials as novel photocatalysts for overall watersplitting, Chem. Commun. (1999) 10771078.

[67] Y. Miseki, H. Kato, A. Kudo, Water splitting into H2 and O2 over niobate andtitanate photocatalysts with (111) plane-type layered perovskite structure,Energy Environ. Sci. (2009) 306314.

[68] J. Yin, Z. Zou, J. Ye, J. Phys. Chem. B 108 (2004) 8888.[69] J. Yin, Z. Zou, J. Ye, Photophysical and photocatalytic activities of a novel

photocatalyst BaZn1/3Nb2/3O3, J. Phys. Chem. B 108 (2004) 1279012794.[70] A. Kudo, S. Nakagawa, H. Kato, Overall wate

photochemical splitting of water for hydrogen production

Documents