Author's Accepted Manuscript

On the engineering part of solar hydrogenproduction from water splitting: Photoreactordesign

Zheng Xing, Xu Zong, Jian Pan, Lianzhou Wang

PII: S0009-2509(13)00582-4DOI: http://dx.doi.org/10.1016/j.ces.2013.08.039Reference: CES11259

To appear in: Chemical Engineering Science

Received date: 7 December 2012Revised date: 16 July 2013Accepted date: 17 August 2013

Cite this article as: Zheng Xing, Xu Zong, Jian Pan, Lianzhou Wang, On theengineering part of solar hydrogen production from water splitting: Photo-reactor design, Chemical Engineering Science, http://dx.doi.org/10.1016/j.ces.2013.08.039

This is a PDF file of an unedited manuscript that has been accepted forpublication. As a service to our customers we are providing this early version ofthe manuscript. The manuscript will undergo copyediting, typesetting, andreview of the resulting galley proof before it is published in its final citable form.Please note that during the production process errors may be discovered whichcould affect the content, and all legal disclaimers that apply to the journalpertain.

www.elsevier.com/locate/ces

Page 1 of 51  

On the engineering part of solar hydrogen production from water splitting: photoreactor design Zheng Xing, Xu Zong, Jian Pan, Lianzhou Wang*  ARC Centre of Excellence for Functional Nanomaterials, School of Chemical Engineering, The University of Queensland, Brisbane, QLD, 4072, Australia E‐mail: l.wang@uq.edu.au   

Abstract  Water splitting under sunlight illumination in the presence of semiconductor photocatalyst is a very promising way to

produce clean hydrogen fuel. Solar hydrogen can be obtained in two routes: photoelectrochemical (PEC) water splitting

based on immobilised photocatalysts in thin films and photocatalytic (photochemical) water splitting based on powder

photocatalysts in slurry system. Over the past several decades, tremendous research work has been devoted to exploring

new semiconductor materials suitable for PEC and photochemical systems and understanding the underlying mechanism

of the water splitting process. However, much less attention has been paid to the design of photocatalytic reaction systems

or reactors, which is indeed critically important for the overall solar energy conversion performance. This paper

summarizes the basic working mechanisms of both PEC and photochemical systems, and gives an overview of a variety of

photoreactor design and development.

Keywords: Photochemistry; Materials; Chemical Reactors; Solar Energy; Hydrogen Production; Reactor Engineering  

1. Introduction  Due to the growing concern of quick exhaustion of traditional fossil fuels including coal, petroleum and natural gas,

mankind have tried to explore renewable energy sources during the past several decades. In particular, this research

enthusiasm was greatly promoted during the energy crisis in mid-1970s. So far, several energy sources are considered to

be renewable energy sources for the future of human beings, including wind energy, tidal energy, geothermal energy, solar

energy, etc. Among these so-called “green energy resources”, solar energy is considered to be the most promising source

Page 2 of 51  

considering the fact that sunlight delivers a power of 1.2×1017 watts continuously (Alic et al., 2010; Morton, 2006; Turner,

1999).

Solar energy can be harvested in a few ways, and generally people try to convert solar energy to other forms of energy,

such as electric energy, thermal energy, chemical energy and so on (Hagfeldt et al., 2010; Kamat, 2007; Lewis and

Nocera, 2006). Efficient utility of solar energy can not only solve the issue of energy shortage, but also release the

pressure from the rising environmental problems, for example, greenhouse gas reduction and water purification(Hoffmann

et al., 1995; Mills and LeHunte, 1997). One of the most studied applications of solar energy is to split water into hydrogen

and oxygen gas on semiconductor photocatalyst.

Hydrogen, which is thought to be the ideal fuel for the future, has a high energy capacity and great environmental

friendliness, and thus reduce our pressure from green gas issue and potential energy crisis. Steam methane reforming

(SMR) is the most widely used method to produce hydrogen so far according to Equation 1. However, shortage of natural

gas and producing of CO2 as the side product are considerable disadvantages of SMR. Production of hydrogen from

carbon-free sources such as water is highly in demand(Guo et al., 2010; Nowotny et al., 2005).

molkJHCOHlOHCH /1.253,4)(2 02224 =Δ+→+ (1)

Since Fujishima and Honda found the photolysis of water into hydrogen and oxygen with a n-type TiO2 semiconductor

photoanode in 1972(Fujishima and Honda, 1972), enormous work has been done in order to achieve photo-assisted water

splitting and realise a sustainable and green way to produce hydrogen. The hydrogen produced in this process, which is

often called solar-hydrogen, is widely recognised to be very promising as clean energy carrier for “hydrogen economy”. In

general, light-assisted water splitting has two types: photoelectrochemical (PEC) water splitting and

photocatalytic/photochemical water splitting, as shown in Figure 1.

 

Page 3 of 51  

 Figure 1. Schematic of (a) photoelectrochemical (PEC) water splitting(Grätzel, 2001) and (b) photocatalytic/photochemical water splitting(Kudo and Miseki, 2008).   In a PEC or photochemical water splitting system, a semiconductor is always involved, which will be excited by incident

photons to produce photogenerated charges. Then the exited semiconductor will react with water and go back to its

original state, and at the same time water molecules are to be broken down and transformed to hydrogen and oxygen. In

the PEC system, an external bias is provided to assist the water decomposition process which involves additional energy

input, while the only energy input in the photocatalytic process is the solar energy. Apparently, the photocatalytic process

is simpler compared to a PEC system without external potential difference to drive the reaction. Considering the fact that

enormous excellent review articles have been published recently in photocatalyst design and photocatalytic /PEC

mechanism areas, in this review, we will focus on the reactor system design of both photocatalytic reaction system and

PEC system, which has received much less attention so far.

2. Background of solar water splitting  

2.1. Photocatalytic water splitting  

2.1.1. Mechanism  Water decomposition is an endothermic reaction with a free Gibbs energy change of ΔG0= +238 kJ/mol as shown in

Equation 2. As a typical “uphill” process, this reaction is not spontaneous in nature, which means that external energy is

needed. The free Gibbs energy change corresponds to 1.23 eV in terms of electrical potential according to Equation 3.

Moreover, if we consider the energy of a single photon, that energy change corresponds to around 1000 nm of wavelength

for a photon, according to Equation 4.

Page 4 of 51  

molkJGgOgHlOH /238),(21)()( 0

222 +=Δ+→ (2)

00 nFEG −=Δ (3)

Where n is the number of electrons transferred in the reaction, F is the Faraday constant

nmnmeVeVE

λ•

=1240)( (4)

 

Unlike insulators or conductors, semiconductors have an energy band structure shown as Figure 2, with a band gap

between the conduction band (CB) and valence band (VB). Upon excitation by external energy source (photic etc.),

electrons in the valence band will be excited to the CB, leaving a “hole” in its original position in VB. In the case of

photoexcitation, the photoexcited electrons and holes exhibited reductive and oxidative power, respectively. In a

photocatalytic water splitting system, electrons in the CB can reduce protons and holes in the valence band to oxidise

water molecules.

 

 Figure 2. Schematic of band structure of semiconductors and water splitting(Chen et al., 2010a).

In order to realize water spitting, the conduction band edge (conduction band minimum, CBM) must lie above the H+/H2

potential (NHE) while the valence band edge (valence band maximum, VBM) must locate below the O2/H2O potential. As

a result, semiconductors are required to possess a band gap over 1.23 eV for the purpose of overall water decomposition.

Page 5 of 51  

Overall photocatalytic water splitting process contains several steps: i) photoexcitation. Photons with sufficient energy hit

the semiconductor so that electrons and holes can be generated in conduction band and valence band, respectively; ii)

charge diffusion. Electrons and holes diffuse from the bulk to the surface reactive sites; and iii) reduction/oxidation

reaction. Charge carriers on the surface react with adsorbed species.

The above-mentioned steps describe the water photolysis process in ideal conditions. In practice, many factors will hinder

the overall water splitting process. From the analysis of whole photocatalytic water splitting process, in the first step,

photon absorption can be a big limiting factor, which is decided by both the light absorption of the whole reaction system

and light absorption properties of the semiconductors. In the second step, one of the greatest obstacles is the fast

recombination of charge carriers. Charge recombination happens very quickly, normally in nanosecond/ picosecond scale.

So the charge carriers need to transfer to the surface of catalysts and react with water molecules before recombination in

the bulk or on the surface. In the last step, one important issue is that semiconductors often have large over-potentials, thus

making the charge transfer between semiconductor and water molecules more difficult. Moreover, the adsorption and

desorption of different species at the semiconductor-solution interface are often dependent on the solution environment.

To address the problem of fast charge recombination, photocatalysts can be modified in various aspects(Liu et al., 2010b).

For example, high crystallinity can enhance the charge diffusion and thus decrease the possibility of charge recombination.

This is because impurities or crystal defects often form recombination centers for electrons and holes, and so highly

crystallized materials will facilitate the charge transfer. Size of catalyst particles also plays a vital role. Generally, the

smaller the crystals, the shorter path charge carriers need to travel to active sites on the particle surface. From this point,

numerous publications focused on synthesis of nano-sized photocatalyst crystals. In addition, usage of sacrificial reagents

(electron scavengers such as AgNO3(Kim et al., 2004; Zong et al., 2011), or hole scavengers such as methanol(Pan et al.,

2011a; Zhou et al., 2010), ethanol(Marschall et al., 2011b; Mukherji et al., 2011a), ascorbic acid(Brown et al., 2012; Han

et al., 2012) etc.) can efficiently promote charge separation by selectively consume charge carriers, that is, electrons or

holes. For instance, addition of hole scavengers in the reaction solution can effectively scavenge photoinduced holes,

leaving electrons to react with water molecules. Schematic of these three aspects is shown in Figure 3.

Page 6 of 51  

 Figure 3. Schematic of the effect of (a) higher crystallinity, (b) smaller crystal size and (c) sacrificial reagents.(Kudo and Miseki, 2008; Liu et al., 2009)

For solving the issue of high over-potential of semiconductors for water splitting, co-catalysts were loaded onto

photocatalysts to facilitate the reactions. Commonly used co-catalysts for water reduction include Pt, NiO, Ru etc. and

those for water oxidation include IrO2, RuO2, Co3O4 etc.. In principle, photoinduced charges will transfer from the

semiconductors to the surface where co-catalyst locates, and then be consumed by adsorbed species.

As many semiconductors are only capable of achieving half of water splitting, either hydrogen evolution or oxygen

evolution, sometimes a H2-evolution semiconductor and a O2-evolution semiconductor are combined to achieve overall

water splitting. As shown in Figure 4, the photoexcited electrons in the H2-evolution semiconductor are used to reduce

protons, and the photoinduced holes in the O2-evolution semiconductor are used to oxidize water molecules. At the same

time, photoexcited electrons in the O2-evolution semiconductor are consumed by the photoinduced holes in the H2-

evolution semiconductor.

Page 7 of 51  

Figure 4. Schematic of Z-scheme photocatalysis(Liu et al., 2009).

 

2.1.2. Evaluation of water photolysis efficiency  If the photocatalysts are tested for water splitting under exactly same conditions, their efficiency can be evaluated directly

by comparing the amount of hydrogen or oxygen gas produced. In such cases, the water photolysis efficiency can be

expressed as moles of gas per unit mass of catalyst per unit time, with the unit of μmol·h-1·g-1 or mmol·h-1·g-1. Using this

method to describe the photocatalysis performance for different catalysts is quite popular, but it is very difficult to

standardize as different research groups may use different photoreactor design, varied reaction solution environment,

diverse gas characterization instruments, etc.. To better explain the photolysis performance of catalysts, researchers often

list amount of catalysts, co-catalyst, reaction solution, sacrificial reagents, reaction system configuration, light sources,

etc..

Introducing of the concept of quantum yield provides another way to present the efficiency. Overall and apparent quantum

yield (QE) are defined as Equations 5 and 6 respectively, and apparent quantum yield is thought to be smaller than overall

quantum yield because incident photons usually outnumber absorbed photons(Chen et al., 2010a).

%100photons absorbed ofnumber

electrons reacted ofnumber yield quantum Overall ×= (5)

100%photonsincident ofnumber

evolved molecules O ofnumber 4

100%photonsincident ofnumber

evolved molecules H ofnumber 2

100%photonsincident ofnumber

electrons reacted ofnumber yield quantumApparent

2

2

××

=

××

=

×=

(6)

Page 8 of 51  

Some materials of the highest quantum efficiency (QE) reported include La-doped NaTaO3 with NiO as co-catalyst (56%

at 270nm, pure water in UV)(Kato et al., 2003), ZnS (90%, aqueous Na2S/Na2SO3, light with λ >300 nm)(Reber and

Meier, 1984), (Ga1-xZnx)(N1-xOx) impregnated with Rh and Cr oxides (5.9% at 420-440nm, pure water)(Maeda et al., 2008;

Maeda et al., 2006a, b; Maeda et al., 2006c).

2.2. Photoelectrochemical water splitting 

2.2.1. Working mechanism of PEC system  When an n-type (p-type) semiconductor is in contact with electrolyte containing the redox couple O2/H2O (H+/H2), the

majority charge carriers, that is, electrons (holes) will flow to the liquid phase due to the difference between the Fermi

level (Ef) of semiconductor and redox potential (Eredox) until an equilibrium state in which the Fermi level is constant at

any point in the system is reached. In the equilibrium state, the semiconductor will have an excess positive (negative)

charge, arising from ionized dopants, while the solution will have an excess negative (positive) charge, and so an electric

field will be formed in the semiconductor near the interfacial region, which counteracts the initial electrochemistry

potential difference between semiconductor and solution. In addition, as a result of interfacial charge flow, the energy

band of semiconductor near the interfacial region will bend depending on the initial difference between the Fermi level of

it is equal to the semiconductor and the electrochemical potential of the redox couple in the solution, which is O2/H2O in

the case of photoanode made of n-type semiconductor(Grätzel, 2001; Lewis, 2005; Walter et al., 2010), shown as Figure 5.

 

Page 9 of 51  

 Figure 5. Band bending of semiconductor when in contact with electrolyte.(Grätzel, 2001; Walter et al., 2010)

In a typical photoelectrode configuration, a photoanode for oxygen evolution is made of n-type semiconductors and a

photocathode for hydrogen evolution is made of p-type semiconductors so that the minority charges carriers are directed to

the solid/liquid interface and react with corresponding redox couple. Take photoanode for example, the interfacial electric

field developed in the n-type semiconductor after reaching equilibrium with the redox couple tends to drive the

photoinduced free minority charge carriers, that is, holes into the electrolyte, and thus O2 gas will be formed.

Based on the semiconductor/electrolyte interfacial energetics described above, photoelectrochemical cells can have some

different configurations. Walter et al.(Walter et al., 2010) summarized schematics of different basic configurations of

photoelectrochemical cells, shown as Figure 6, which are assumed to possess a back-to-back “wireless” device design

established by Nozik first(Nozik, 1977). Specifically, PEC cells can be built in four configurations: direct combination of

a single semiconductor photoelectrode along with a counter-electrode that is normally made of metal (Figure 6a);

connection between a n-type semiconductor and a p-type semiconductor in series (Figure 6b); coupling of a single

photoelectrode and a photovoltaic (PV) device, in which PV is used to provide extra bias (Figure 6c); joint of two PV

devices in series (Figure 6d).

  

Page 10 of 51  

Figure 6. Schematics of different PEC cell configurations: (a) a single n-type semiconductor photoanode PEC in contact with metal cathode back; (b) a dual band gap PEC with a n-type and p-type semiconductor photoelectrodes connected in series; (c) a n-type photoanode connected in series with a PV which has a metal cathode for hydrogen evolution; and (d) two PV connected in series(Walter et al., 2010). The PV device in the PEC configurations mentioned above can be replaced by a dye-sensitized solar cell (DSSC), which

essentially also functions as an extra bias provider. Considering the quick advancement of DSSC lately(Bai et al., 2012;

Wu et al., 2011a; Wu et al., 2011b), DSSC-contained PEC provides a promising development direction. In fact, the PEC

cells do not need to be assembled in such back-to-back style and in many cases they possess very different forms: the

external power sources, that is, PV devices, DSSCs or power sources from the grid can be placed in external circuit; the

cathode and anode can be immersed into the same vessel containing electrolytes or into different electrolytes, which is

sometimes used to provide additional voltage or so-called “chemical bias”(Minggu et al., 2010; Selli et al., 2007). More

and more new PEC forms began to show up very recently. Seger et al. introduced a new concept of PEC cell which

combined hydrogen evolution at the cathode and organic degradation at the photoanode(Seger et al., 2012). In their design

of membrane electrode assembly (MEA), photoanode made from Aeroxide P-25 TiO2 and cathode made from carboxy

functionalized multiwalled carbon nanotubes (C-MWNT) deposited with Pt were hot pressed onto Nafion 117 membrane.

Organic chemicals such as methanol was photodegraded to CO2 and protons at the photoanode, and then protons travelled

through the Nafion membrane to the cathode, where they were reduced to H2, functioning similarly as a fuel cell (Figure

Page 11 of 51  

7). Another new PEC design was come up with by Marschall et al. in which no chemical or electrical bias was needed to

drive hydrogen evolution process(Marschall et al., 2011a). As shown in Figure 8, a proton exchange membrane (PEM)

such as commercial Nafion®, FKE® membrane (fumatech) and home-made sulfonated polyethersulfone polymer

membrane was exploited to support and separate the carbon coated Degussa TiO2 P25 photoanode and the Pt cathode. In

this PEC cell, both electrolyte for applying chemical bias and external power source for applying electrical bias are absent.

 

 Figure 7. The schematic of (a) PEC cell and (b) overall reaction system designed by Seger et al.(Seger et al., 2012)  

Page 12 of 51  

 Figure 8. The schematic of PEC without chemical or electrical bias(Marschall et al., 2011a).  

2.2.2. Evaluation of PEC performance  In a typical three-electrode PEC cell containing a working photoelectrode, a counter-electrode and a reference electrode,

the system solar-to-chemical conversion efficiency ( ) can be calculated according to Equation 7(Walter et al., 2010).

in

ex

PVVJ )23.1( −

=η (7)

where J is the current density measured in the external circuit, Vex is external potential applied between photoanode and

photocathode and Pin is power density of light coming to the PEC cell.

In such configuration, evolved oxygen and hydrogen need to be separated in order to suppress the back reaction. Only so

can accurate efficiency be obtained.

Page 13 of 51  

For a dual band gap p/n semiconductors cell, the performance of a single photoelectrode can be evaluated according to

Equation 8(Walter et al., 2010) from the J-V curve of it.

ocsc

mmmm

in

mm

VIVIffVJ

PVJ

××

=×=×

= P mη (8)

where η is the photoelectrode efficiency, Jm is the current density at the maximum power point, Vm is the voltage at the

maximum power point, Pin is the input powder of light source, Pm is the maximum powder density, ff is the fill factor, Im is

the current at the maximum power point, Isc is the short circuit current and Voc is the open circuit voltage.

Here Isc and Voc reflect the thermodynamic potential of the photoelectrode for the corresponding half water splitting

reaction, while the fill factor is often associated with charge recombination. Equation 8 only described the characteristics

of a single photoelectrode. To decide the overall efficiency of an n/p PEC cell, J-V curves of both photoanode and

photocathode need to be plot in the same figure, and then the intersection point will give the actual operation point with

the maximum power.

3. Photocatalytic water splitting reactors  For a photochemical water splitting system, the photocatalyst will be mixed with water forming a homogeneous or

heterogeneous slurry suspension. This suspension will absorb photons from an artificial light source or natural solar light,

and then react between solid-liquid-gas three-phase interfaces. Such a slurry-based photocatalytic reactor is often exploited

for photochemical water splitting, which is called “semiconductor-particles-based” device by Tachibana et al.(Tachibana

et al., 2012). From a theoretical point of view, a reliable photoreactor should be capable of effectively absorbing incident

light with minimal photonic losses and facilitating photocatalytic reactions. On the other hand, from the angle of practical

application or probable scale-up to industrial usage, technical challenges and cost related issues are equally important to be

considered in the construction of new reactor design(Huang and Yao, 2011). A number of publications reviewed the

different types of photoreactor design and analysis of important factors to consider for building an efficient photoreactor in

the area of photocatalytic water treatment application, but only a few focused on water splitting purpose(Alfano et al.,

2000; Bahnemann, 2004; Bahnemann, 2000; Braham and Harris, 2009; Huang and Yao, 2011; McCullagh et al., 2011).

Different from water treatment photoreactors, good sealing is needed for purpose of protecting reactant from air or/and

Page 14 of 51  

avoiding hydrogen loss for photochemical water splitting photoreactors. Here we will look into the design for

semiconductor-particles-based photoreactors.

A simple photocatalytic water splitting system set-up can be constructed from connecting a Pyrex reactor with gas

collector(Khnayzer et al., 2011; Mangrulkar et al., 2010; Mangrulkar et al., 2012). Figure 9 shows a typical configuration

of this type. After all reactants are transferred to the tubular Pyrex glass reactor, a condenser is applied on the top of the

reactor and connects the reactor with a gas collector. After reaction, the gas collector was detached from the reactor and

then connected to a Gas Chromatography (GC) to detect the products.

 

 Figure 9. A simple photoreactor set-up constructed by connecting Pyrex reactor with a gas collector(Mangrulkar et al., 2012).  

3.1. Batch­type photoreactor  

So far, the dominant type of photoreactor is batch-type reactor. Figure 10 shows the schematic of a typical batch-type

reactor(Huang and Yao, 2011). In such a reactor scheme, the reaction slurry suspension is placed in a reaction tank (made

of stainless steel, Pyrex, quartz, etc.), where a magnetic stirrer adequately stirs the slurry and thus stops the catalyst

particles to deposit and accumulate. There is often water jacket surrounding the tank, allowing cooling water running

around the reactor and thus keeping the reaction process in a certain temperature. Artificial light sources will be irradiated

upon the top quartz window of the reactor.

Page 15 of 51  

Besides the typical batch-type photoreactor showed above, a complete photocatalytic water splitting system will also

integrate the following parts: light source, evacuation system, sample/product collecting device, and gas detection

instrument. We will talk about each part in detail.

Figure 10. Schematic of a typical batch-type photoreactor built from stainless steel vessel and quartz window on the top,

water jacket is placed around the vessel with cooling water to keep the temperature constant(Huang and Yao, 2011).

3.1.1. Batch-type reaction tank

3.1.1.1. Geometric set-up of photoreactor Unlike conventional chemical reactors, physical geometry is critically important for photoreactors because it determines

the photon collection efficiency of the overall photochemical system. Photoreactors with various shapes are developed in

different trials. Some researchers have done radiation absorption calculations and simulations for different

photoreactors(Alfano et al., 2000; Brandi et al., 2000; Huang et al., 2011; Romero et al., 1997). However, it is very

difficult to analyze the geometric design of photoreactors as the uneven light intensity distribution inside the reaction tank

and complexity of different physical factors involved. In particular, light distribution determines the local photon

absorption rate and thus photochemical reaction rate(Quan et al., 2004; Yang et al., 2005). Light distribution can be

affected by reflection and scattering of the reaction tank and the powder catalyst, and sometimes blocking of light by

reactants or supporting materials. In addition, the position of the light source poses great influence to the light distribution.

Page 16 of 51  

For example, in a typical cylindrical photoreactor, light intensity distribution can be totally different depending on whether

the lamp is placed on the top or around the perimeter. Inside the photoreactor, there are three possible regions: strong

illumination region (close to the lamp), weak illumination region (away from the lamp) and dark region, of which

photocatalytic reaction can only happen in the first two regions (Huang et al., 2011). So an optimized geometrical

configuration is very important for both lab-scale study and potential scale-up applications. For lab-scale study, cylindrical

reaction vessel is suggested as it allows relatively uniform stirring, while for potential scale-up applications, a geometry

with light concentrating function such as compound parabolic concentrator (CPC) is preferred because in practical

conditions attainment of higher light intensity under direct sunlight is essential for the overall efficiency. This will be

discussed in detail in the standardization part later.

3.1.1.2. Tank materials The tank can be built from different materials including stainless steel(Huang and Yao, 2011), Pyrex(Chen et al., 2011;

Escudero et al., 1989; Ishikawa et al., 2002; Khnayzer et al., 2012; Kim et al., 2011; Lo et al., 2010; Sabate et al., 1990;

Yu et al., 2011) and quartz(Marschall et al., 2011b; Mukherji et al., 2011b). The materials of the tank play an important

role in affecting the photocatalysis efficiency as it determines the light transmission. Quartz is certainly ideal for optical

transmittance, but considering its high cost, it is not very frequently used. Stainless steel could effectively block diffuse

light from surrounding environment and thus control the illumination area relatively precisely. However, it is not proper

for scale-up application due to the cost issue and inability to transmit light. Pyrex is a type of borosilicate glass, originally

introduced by Corning Incorporated. It features low thermal expansion coefficient and refractive index, and relative lower

cost compared to fused silica. Due to its good physical and optical properties, Pyrex is a very popular photoreactor

material. But for accurate photocatalytic reaction analysis purpose, it has the drawback of allowing unexpected light to go

into the system.

3.1.1.3. Window materials Window materials are critically important for illumination condition in photocatalytic system, so a material with high

photon transmission property and low cost is highly desirable. Almost all photoreactor windows exploited are made of

quartz, or fused silica, to date. Fused silica (SiO2) is a synthetic molten, amorphous quartz glass, with very high optical

Page 17 of 51  

transmission for UV(Huang and Yao, 2011). Despite of its good optical property, high price of fused silica limited its

application in potential future large-scale usage.

In order to find alternative low-cost window material with good light transmittance to replace fused silica, Huang et

al.(Huang and Yao, 2011) tested 8 different window materials including fused silica, Pyrex glass, Aclar, Kynar PVDF,

Dupont Teijin Mylar, PET, PVC and hot mirror and found that Aclar is actually more efficient than Pyrex. Their findings

provided potential alternative window materials.

3.1.2. Cooling system Temperature is an important factor in the photocatalytic reaction. Due to photonic activation, the true activation energy for

photocatalysis is negligible but the apparent activation energy is often affected by some factors (Herrmann, 2005). At

higher temperature, the viscosity of solution will decrease, thus boosting the detachment of product bubbles from the

catalyst powder surface (Huang and Yao, 2011). Moreover, the detrapping of photoinduced charge carriers will be

improved at elevated temperature (Hisatomi et al., 2009). However, at the same time, the exothermic adsorption of

reactants at the catalyst surface will be hindered (Herrmann, 2005). On the contrary, lower temperatures will favor the

adsorption of reactants and disfavor the desorption of photoproducts.

A cooling system is essential for the photoreactor, especially in the catalyst performance comparison study, to keep the

reaction proceeding in a constant temperature. A water jacket is often located outside the photoreactor with recycling cool

water running through it. Some experiments are done in an environment of 25 °C (Chen et al., 2011; Priya and Kanmani,

2009).

3.1.3. Light source Ultimately, photocatalysts are designed to work under illumination of sunlight. So a solar simulator is often used in

photocatalytic reactors to produce similar spectrum like sunlight. For example, Huang et al. (Huang and Yao, 2011) used

a Diversion beam solar simulator (Newport/Oriel Corp.) equipped with a 1000W ozone free xenon lamp, and an air mass

(AM) 1.5 global filter. Figure 11 shows the comparison between a standard solar spectrum and the output spectral

distribution of an Oriel AM 1.5 Direct Simulator (Newport Corp.). As solar simulator models manufactured by different

companies may vary, international standards have been developed to quantify the performance from three aspects: spectral

match, non-uniformity of irradiance and temporal stability (http://www.newport.com/images/webdocuments-

Page 18 of 51  

en/images/Solar_Industry-Solar_Simulation.pdf). There are three widely recognized international standards: the

International Electrotechnical Commission (IEC) 60904-9 Edition 2 (2007); the Japanese Industrial Standards (JIS) C

8912 (1998); and the American Society for Testing and Materials (ASTM) E927-05 (2005). These international standards

classify the above-mentioned three aspects of solar simulators into three classes: A, B, or

C(http://www.eyesolarlux.com/Solar-simulation-ASTM-IEC-JIS.htm). As can be seen from the top AAA classes with the

highest specifications in Table 1, varied standards show minimal difference while the testing methods can differ from one

another. However, many manufacturers such as Newport Corp. and Yamashita Denso Corp. can produce AAA class solar

simulators which can meet all the three standards (http://www.newport.com/Oriel-Sol3A-Class-AAA-Solar-

Simulators/842468/1033/info.aspx#tab_Overview; http://www.yamashitadenso.co.jp/english/product03.html).

Figure 11. A comparison between the output spectrum of a typical Oriel AM 1.5 Direct Simulator (Newport Corp.) and the standard solar spectrum according to the ASTM (the American Society for Testing and Materials) E891(http://assets.newport.com/webDocuments-EN/images/12298.pdf). Standards IEC 60904-9 (2007) JIS C 8912 (1998) ASTM E927-05 (2005) Spectral match 0.75-1.25 0.75-1.25 0.75-1.25 Non-uniformity of irradiance 2% <±2% 2% Temporal Instability <2% <±1% 2% Table 1. The specifications of AAA classes according to different international standards(http://www.newport.com/Oriel-Sol3A-Class-AAA-Solar-Simulators/842468/1033/info.aspx#tab_Overview).

Besides solar simulator, other light sources (lamps) have also been used since the very beginning of photocatalyst

research, such as xenon lamp(Chen et al., 2011; Khnayzer et al., 2012; Sabate et al., 1990; Yu et al., 2011), mercury

lamp(Escudero et al., 1988; Escudero et al., 1989), halogen lamp(Lo et al., 2010), etc.. As different light sources will

Page 19 of 51  

produce different spectrum, the choice of light source is important. Coupling between light spectrum and the absorption

properties of semiconductor can promote the water decomposition efficiency.

The light intensity of illumination sources can be monitored by some instruments: radiometer (Goldilux GLP-1

radiometer, model 70235)(Lo et al., 2010), Lumen meter (Goldilux GRP-1 70234)(Chen et al., 2011), calibrated

photospectrometer (OMS-100-UV/VIS, Newport/Oriel Corp.) connected a Sony CCD sensor(Huang and Yao, 2011),

Oxalic-uranyl actinometry and a laser power meter (Newport Corp., model 820)(Sabate et al., 1990) and so on. In most

cases, the light intensities in front of the reactor are reported to better describe the photocatalytic water splitting reactor

set-up.

3.1.4. Evacuation system Before the water splitting reaction proceeds, the overall system needs to be evacuated so that all the air can be removed.

A way to do this is to purge the whole reaction system with inert gas for some time, such as ultra-pure Argon

gas(Escudero et al., 1988; Huang and Yao, 2011; Lo et al., 2010; Yu et al., 2011) or nitrogen gas(Priya and Kanmani,

2009). In the water splitting experiment of Chen et al.(Chen et al., 2011), they first heated the solution to 50 °C and

evacuate, and purged with Ar and re-evacuate for several times subsequently. To further eliminate the effect of air, the

amount of remaining air was measured by gas chromatography.

Another way is to connect the reaction vessel to a closed gas circulation and evacuation system(Ishikawa et al., 2002;

Ishikawa et al., 2004). Through the evacuation by an external pump, the air contained in the reaction system can be

effectively driven out.

3.1.5. Gas detection After the hydrogen or/and oxygen gas evolved from the reaction slurry, the gas-phase products need to be collected and

analyzed. The product can easily be collected in an inverted container so that its volume can be determined by the water

displacement(Huang and Yao, 2011). However, to accurately check the product amount and species, gas chromatography

(GC) is often employed. GC is often equipped with thermal conductive detector (TCD) and molecular sieve packed

column, with inert gas (Ar etc.) running through it as the carrier gas(Chen et al., 2011; Escudero et al., 1988; Escudero et

Page 20 of 51  

al., 1989; Huang and Yao, 2011; Khnayzer et al., 2012; Kim et al., 2011; Lo et al., 2010; Priya and Kanmani, 2009; Yu et

al., 2011). Besides GC, quadrupole mass spectrometer is sometimes exploited to obtain the partial pressure of the product

and converted moles(Marschall et al., 2011b; Mukherji et al., 2011b).

3.1.6. Overall system set-up Figure 12 shows two different kinds of batch-type photoreactor overall system set-up. Chen et al. used a Pyrex reaction

cell, illuminated by a Xe lamp (λ > 400 nm) from the side and a 254 nm UV lamp placed in the center of the reactor(Chen

et al., 2011), while Mukherji et al. performed water splitting in a quartz reaction cell with a sun simulator integrated with

AM 1.5 filter(Mukherji et al., 2011b). Both systems are desecrated via purging with Ar gas, but Chen et al. exploited GC

to detect gas produced while Mukherji et al. measured the amount of evolved hydrogen with quadrupole mass

spectrometer. The overall system set-up can be significantly different among so many research groups all over the world,

depending on the set-up of evacuation system, sampling system, gas detection system etc. As a result, there exists great

difficulty in the standardization of photocatalyst evaluation.

Figure 12. Schematics of two different kinds of batch-type photoreactor overall system set-up: (a) Pyrex reactor is illuminated by Xenon lamp from the side and UV lamp in the center, while gas is detected by GC; and (b) quartz reactor is

Page 21 of 51  

illuminated by sun simulator and gas is detected by quadrupole mass spectrometer (Chen et al., 2011; Mukherji et al., 2011b).

3.2. Modification of batch­type and other type photoreactors   In spite of the prevalence of batch-type photoreactors in the photocatalytic water splitting area, it has some limitations. As

described before, thorough stirring of the slurry in photoreactor is essential for the reaction to proceed, which stops the

photocatalyst powder from settling and accumulating, but in large-scale applications, it is not feasible to use magnetic

stirring due to the cost issue. For large-scale applications, a passive method is more desirable than the magnetic stirring

method to provide the photocatalyst uniform irradiation condition. Huang et al.(Huang and Yao, 2011) proposed two ways

to realize passive powder-liquid mixing method: coating catalyst powder onto a packing materials with high surface area

that fills the photoreactor; configuring the locations of inlet and outlet ports of the photoreactor to create enough

turbulence in the flow to prevent the catalyst powder from sinking.

Generally, the overall water photolysis rate is determined by three factors except the intrinsic photocatalytic efficiency of

catalyst: light absorption of the photoreactor, reverse reaction on the catalyst surface and transfer of products from catalyst

surface to gas phase(Escudero et al., 1989). Light absorption properties of photoreactors rely on their geometric shape and

the building materials. For overall water splitting, hydrogen and oxygen are produced simultaneously on the catalyst

surface, so the reverse reaction, that is, recombination of hydrogen and oxygen, will inevitably happen, which greatly

deteriorate the photocatalysis efficiency. The last factor is often not paid enough attention to. In batch reactors,

photoproducts tend to have difficulty in moving from the liquid to the gas phase because of the small gas-liquid interfacial

area. In addition, build-up of photoproducts in the gas phase can hindered the photocatalysis process as the driving force of

photoproduct transfer is weakened.

In order to address the issues discussed above about batch-type reactors, researchers have been trying to modify it or to

design new type of photoreactors.

3.2.1. Twin reactor 

Conventionally, for a Z-scheme photocatalytic water splitting material system involving two semiconductors, the H2-

evolving catalyst and O2-evolving catalyst are mixed in a single reactor to achieve water photolysis. In such configuration,

Page 22 of 51  

back reaction is a big concern as mentioned before, and separation of hydrogen will cause extra expense. To overcome this

drawback, a twin reactor or membrane reactor was developed, shown as Figure 13. In the twin reactor system, H2-evolving

semiconductor and O2-evolving semiconductor are placed in different compartments of a connected twin-reactor separated

by a Nafion membrane (proton exchange membrane). When Fe2+/Fe3+ are used as the redox couple, the Nafion membrane

with a thickness of 178 μm will allow Fe2+/Fe3+ ions’ counter penetration. Before usage, the membrane needs to be cleaned

by different acid and alkali solutions. The produced H2 and O2 are collected separately by an online sampling loop (1 mL)

by switching valves alternately. To avoid cross contamination, the gas lines are evacuated and purged with Ar between

two sampling processes(Lo et al., 2010; Yu et al., 2011).

Figure 13. Schematic of twin reactor system: each compartment has a volume of 180 mL, and the two compartments are connected through a Nafion membrane(Lo et al., 2010).

3.2.2. Batch­recycle reactor 

To scale up water splitting photoreactors, batch-type reactors have been modified to batch-recycle reactors(Huang and

Yao, 2011; Oralli et al., 2011; Priya and Kanmani, 2009). A general configuration of batch-recycle reactor is depicted in

Figure 14a, which circulates the reaction solution between a photocatalytic reactor and a tank containing more solution by

a pump. Batch-recycle reactors must be designed to ensure enough turbulence in the flow in order to avoid particle

settling. In a typical batch-recycle reactor system developed by Huang et al.(Huang and Yao, 2011) shown in Figure 14b, a

pump was employed to circulate the photolyte and photocatalyst powder from a three-neck flask to one port of the

photoreactor. The photoreaction vessel was made up of stainless steel and machined to taper-shaped inside with a slope of

10-15 degree. The reaction slurry entered the photoreactor from the edge and exited at the center of bottom (Figure 15).

This configuration provided a method to realize passive mixing, allowing high degree of mixing and thus hindering the

Page 23 of 51  

formation of dead zones and particle sedimentation. Another batch-recycle system established by Priya et al. is shown in

Figure 14c. In their set-up, N2 was continuously fed to the photoreactor through a sieve plate at the bottom of the

photoreactor and the solution was circulated by a pump underneath the reactor. The whole system is placed in direct

sunlight, and gas products are collected from a port in the photoreactor and analyzed by GC.

Figure 14. Schematic diagrams of (a) general configuration of batch-recycle photoreactors(Oralli et al., 2011); (b) batch-recycle system build from three-neck flask and taper-shaped photoreactor made of stainless steel(Huang and Yao, 2011); and (c) batch-recycle system with continuous N2 feed in the absence of solution tank(Priya and Kanmani, 2009).

Page 24 of 51  

Figure 15. Detailed constructin of batch-recycle photoreactor shown in Figure 16b: (1) photoreactor; (2) gasket; (3) flange; (4) hot mirror; (5) inlet port; and (6) outlet port.(Huang and Yao, 2011)

3.2.3. Continuous annular photoreactors 

Continuous annular photoreactors were also developed for better addressing the detachment problem of gas bubbles from

the catalyst particles(Escudero et al., 1990; Escudero et al., 1988; Escudero et al., 1989; Esplugas et al., 1987). Revealed as

Figure 16, a continuous annular photoreactor consist of an annular shaped reactor with a lamp located in its center. The

reactor is placed in an inert housing filled with nitrogen to separate the system from surrounding air outside. Ar gas is

continuously bubbled through the reaction slurry, keeping the suspension in good mixing state. Part of the generated gas is

introduced to a gas valve and subsequently analyzed by GC. This design greatly increases the liquid-gas interfacial area,

thus facilitating desorption of photoproducts from catalyst surface. The drawback of this reactor is that the interior of the

annular reactor can accept more photons than the exterior perimeter.

Figure 16. Schematic diagram of continuous annular photoreactor.(Escudero et al., 1990)

3.3. Practical trials  Even though photocatalytic water splitting has been researched for several decades, it is still mainly confined in lab-scale

study and there are very limited amount of review literatures available regarding its reactor design. Comparatively

Page 25 of 51  

speaking, the reactor study of photoreactors for water purification is more mature and many researchers have looked into

industrial application. Braham et al.(Braham and Harris, 2009) summarized different photoreactor configurations to date

including parabolic trough reactor (PTR), compound parabolic collector (CPC), inclined plate collector (IPC), double skin

sheet photoreactor (DSS), rotating disk reactor (RDR) etc., mainly for water treatment purpose. Although photoreactors

for water splitting and water purification are different, Braham’s work provides some very useful hints for the design of

water photolysis reactors.

Jing et al.(Jing et al., 2010) tried to use a compound parabolic concentrator (CPC) for hydrogen production. In their

scenario, a CPC (Figure 17) was coupled with an inner-circulated reactor in order to achieve higher sunlight energy

intensity. The whole photolcatalysis system was placed outdoor under direct illumination of sunlight (Figure 18). For the

capture of maximum photons from sunlight, the aperture of the CPC was set to be perpendicular to the incident light as far

as possible. The CPC design is thought to enable solar rays to hit the complete perimeter of the round receiver, rather than

just the “front” of it. The fluidization of the slurry was carefully adjusted to reach a Reynolds number between 10000 and

50000 to ensure fully turbulent flow and good dispersion of photocatalyst particles so that sedimentation and accumulation

of catalyst were avoided.

Figure 17. Geometric profile of CPC(Alfano et al., 2000).

Page 26 of 51  

Figure 18. Photocatalytic hydrogen production reactors built from CPC(Jing et al., 2010). Distribution of catalyst particles in the suspension is an important factor determining the light absorption in the reactor.

Through Jing’s testing, distribution depended on slurry pressure gradient, catalyst concentration, slurry velocity etc.

When the loading amount of the CdS photocatalyst, Na2SO3 and Na2S sacrificial agents is 1.0 g/L, 0.1 mol/L and 0.1

mol/L, respectively in a total system volume of 11.4 L, a maximum hydrogen production rate of 1.88 L/h can be obtained

with an efficiency of 0.47%(Jing et al., 2009). Despite of its low efficiency, Jing’s trial showed the potential large-scale

practical application of photocatalytic water splitting reactors.

3.4. Standardization of photocatalytic water splitting evaluation 

So far, there is no standard testing system for semiconductor-based photocatalytic water splitting, which means that it is

impossible to directly compare the performance of photocatalysts developed by different research groups. To improve the

standardization in lab-scale evaluation of photocatalysts, several aspects which need to be paid attention to:

(1) Light source: as the ultimate goal of photocatalytic water splitting research is to produce hydrogen gas under

sunlight illumination, the best light source is solar simulator coupled with an air-mass 1.5 filter (AM1.5), which

produces light beams with similar spectrum of sunlight and light intensity of around 1000 W/m2. AAA class

solar simulators can produce very close approximation of solar spectrum, which we think are good enough for

the photocatalysis evaluation purpose at this stage. As mention before, the light intensity in front of the reactor is

usually measured rather than the real intensity of light which goes into the reactor. Considering this issue,

Maschmeyer et al. pointed out that this measurement can be done easily and reproducible by actinometry, which

has been well-established(Kuhn et al., 2004; Maschmeyer and Che, 2010; Parker, 1953).

Page 27 of 51  

(2) Reactor configuration: for half reaction of water splitting, a normal batch-type reactor with cylindrical shape is

suggested as it allows relatively uniform stirring. Light source can be located on the top or around the perimeter,

but the “window” of the photoreactor needs to be fixed so that the area through which light goes into the

photoreactor is constant. The best choice of window material is quartz and the other parts of photoreactor should

be masked. Volume of the photoreactor should be proper so that reasonable amount of reaction solution can be

placed in the reactor and some free space is reserved to avoid high pressure at the same time. The volume may

vary from 100 ml to 1 L. Temperature controlling equipment must be applied to ensure that the reactor is running

at a constant temperature.

(3) Amount of reactants: the amount of catalyst should be appropriate for the volume of reaction solution and the

overall photoreactor. For example, 100 mg of catalyst is often exploited when a total volume of reacting solution

is around 300 ml (Liu et al., 2010a; Pan et al., 2011b; Wang et al., 2009; Zong et al., 2011). Moreover, the

amount of co-catalyst and sacrificial reagents must be strictly controlled because excessive cocatalysts may block

the photon absorption by the semiconductor. In addition, the best performance of different semiconductors may

be achieved with different co-catalysts, which makes it difficult to standardize the reactants. For example, Pt is a

well-known cocatalyst for hydrogen production but some semiconductors exhibit better performance with other

metals (e.g., Ru and Rh)(Hara et al., 2003; Maeda, 2011; Maeda and Domen, 2011). Maschmeyer et al.

addressed that the system must be operating in a “non-diffusion-limited” regime, in which twice of the catalyst

loading results in twice of the reaction rate(Maschmeyer and Che, 2010).

(4) Vacuum: high vacuum level is crucial for water photolysis evaluation. Considerable amount of remaining air

inside the photoreactor will not only greatly affect the detection of gas products, but also influence the inner

pressure and thus the transfer of photoproducts to the gas phase(Maeda et al., 2006c; Sayama and Arakawa,

1997). In this regard, vacuum pump is preferred rather than gas purging as the former is more effective in

producing an air-free environment.

As the purpose of lab-scale evaluation is effective materials screening, the standardization of photoreactor is of vital

importance as it allows researchers to cross-compare different photocatalysis systems. For large-scale applications, since

the research in practical water splitting reactor is still in its very early stage, standardization seems not to be very essential

at the moment. Here we will give some aspects to consider in terms of standardization of large-scale photoreactors:

Page 28 of 51  

(1) Illumination: The photoreactor should be placed outdoor under direct sunlight. However, the problem with direct

sunlight is that the light intensity strongly depends on the weather, geographical position, time in the day,

geometric shape of the photoreactor etc., and so using similar method as mentioned in the lab-scale study to

detect the light intensity is very important.

(2) Geometric shape of photoreactor: Enlargement of photoreactor will greatly increase the fabrication cost, so the

geometric shape of the photoreactor needs to be carefully designed in order to maximize the light usage

efficiency. Photoreactor with light concentrating function like CPC should be exploited.

(3) Reactor construction material: Different from lab-scale study, for large-scale application, the construction cost

should be taken into serious consideration. More cost effective materials like Aclar polymer film as Huang et

al.(Huang and Yao, 2011) suggested to replace quartz should be identified to construct the relative large reactor

in large-scale application.

4. PEC reactors   For photoelectrochemical water splitting, the photoreactor will involve photoelectrodes made from semiconductors,

electrolytes and so on. The apparatus is often termed as photocell. A few good reviews regarding photoreactors or

photocells for PEC have been published in recent years(Carver et al., 2012; James et al., 2009; Minggu et al., 2010). In

particular, Minggu et al.(Minggu et al., 2010) systematically summarized the reacting vessels for PEC, from the working

mechanisms to the practical configurations, and they also presented trials of potential large scale production. This review

will briefly sum up different photocell configurations.

4.1. Fabrication of photoelectrodes  Photoelectrodes are generally made by depositing a thin layer of semiconductor materials onto the substrate with a

conducting top surface. The substrate can be transparent conducting glass such as fluorine-doped tin oxide (FTO)(Karuturi

et al., 2012b; Luo et al., 2012; Maeda et al., 2011; Paracchino et al., 2011) or metal piece(Higashi et al., 2012; Park et al.,

2006). Semiconductor materials can be deposited simply by doctor blading method(Tetreault et al., 2011), new

technologies have been developed to synthesize better quality photoelectrodes in which the morphology and other

properties of the semiconductor materials can be precisely controlled, including anodization(Li et al., 2011; Park et al.,

2006), radio frequency (RF) magnetron sputtering(Vidyarthi et al., 2011), hydrothermal growth method(Luo et al., 2012),

electrophoretic deposition(Higashi et al., 2012), atomic layer deposition(Paracchino et al., 2011) etc.. A better quality

Page 29 of 51  

photoelectrode can be fabricated via improved ohmic contact processing and proper insulation(Hou et al., 2011; Miller and

Rocheleau, 1999; Minggu et al., 2010; Tseng et al., 2011; Varner et al., 2002). After deposition of semiconductor on

photoelectrode, a blank area of the substrate is often kept undeposited so that a copper wire can be attached to the

photoelectrode with silver glue. Epoxy resin can be used to cover the photoelectrode except the photoactive area and the

copper wire can even be protected by glass tube. Figure 19 shows the fabrication of photoelectrode(Minggu et al., 2010).

Figure 19. Fabrication of photoelectrode(Minggu et al., 2010). Photoelectrodes can be fabricated from a single semiconductor or stack of several different semiconductor layers or even

homogeneous hybrid materials. For example, semiconductors with different band gaps can be deposited onto one substrate

layer-by-layer, and thus enable the photoelectrode to absorb different parts of sunlight spectrum(Graetzel and

Augustynski, 2005). Some semiconductor layers can also be deposited onto the top of others to protect the unstable

semiconductors inside(Paracchino et al., 2011).

4.2. Different types of photocells  The first case for photoelectrochemical water splitting was realized by Fujishima and Honda(Fujishima and Honda, 1972)

in 1972 as mentioned previously. Their PEC system is simply composed of a photoanode made from n-type TiO2 and Pt

counter-electrode, shown as Figure 20.

Page 30 of 51  

Figure 20. Photoelectrochemical water splitting system developed by Fujishima and Honda in 1972(Fujishima and Honda, 1972). 1: photoanode made from TiO2; 2: Pt counter-electrode; 3: ion exchange membrane; 4: external bias; 5: potentiometer. Since Fujishima and Honda’s pioneering work, tremendous trials have been done in order to develop PEC cells with

improved efficiency, from primitive forms to very sophisticated configurations. Similar to photoreactor of slurry-based

water splitting, the physical design played a vital role in determining the final performance. Here we have divided all PEC

cells into two categories: open photocells and photocells with gas separation and collection.

4.2.1. Open photocells  To measure the solar to current conversion efficiency, a reference electrode is often used. An open photocell with different

electrodes, that is, working electrode, counter electrode and reference electrode, is the most basic configuration. Normally,

the three electrodes are immersed in the electrolyte in an open vessel and the potential of the working electrode is

controlled by a potentiostat. This type of photocell is called three-electrode configuration, and is probably the most

commonly used one in lab-scale investigations(Higashi et al., 2012; Karuturi et al., 2012a; Luo et al., 2012; Park et al.,

2006; Tseng et al., 2011). Figure 21 shows a typical configuration of three-electrode PEC used by Maeda et al.(Maeda et

al., 2011). In their experiment, SrNbO2N composed the working electrode, and Pt wire composed the counter electrode

while saturated Ag/AgCl reference electrode was exploited. The electrolyte was Na2SO4 solution (pH=6). Evolved gases

can be analyzed using GC. The photocell vessel can be made from quartz or Pyrex glass while its shape may range from

square to semi-round or round in order to better utilize incident light(Kelly and Gibson, 2006, 2008; Maeda et al., 2011;

Miller and Rocheleau, 1999).

Page 31 of 51  

Figure 21. Three-electrode PEC configuration(Maeda et al., 2011). The distances between electrodes are very important, so researchers are always trying to keep the distances constant during

photoelectrochemical measurements. For such a purpose, some have developed PEC cells with fixed ports for different

electrodes, shown as Figure 22. PEC vessel made of quartz in Figure 22a possesses a circular shape with three ports at the

top and on both sides(Dresselhaus and Thomas, 2001). The rectangular cell in Figure 22b is designed in such a way that

there are two ports for reference electrode and counter electrode located symmetrically at the bottom while the working

electrode is inserted to the cell from the top opening(Kamat, 2007).

Page 32 of 51  

Figure 22. PEC cells with dedicated ports for different electrodes: (a) circular cell with ports at the top and on both sides(Dresselhaus and Thomas, 2001); and (b) rectangular cell with two ports located symmetrically at the bottom(Kamat, 2007).

4.2.2. Photocells with gas separation and collection  As in the photoelectrochemical water splitting process both hydrogen and oxygen gas will evolve, gas separation is quite

essential considering the issue of gas collection and safety hazard. So far, many PEC modules with gas separation

functionality have been proposed.

4.2.2.1. Ion­permeable membrane separated PEC cells  One way to separate gas is to place an ion-permeable membrane between the anode and cathode, or to build two chambers

for anode and cathode interconnected by the membrane. The membrane will allow ion exchange between the two

compartments of the photocell but hydrogen and oxygen can be separately collected. As shown in Figure 23a, Selli et al.

build a PEC cell of this type(Selli et al., 2007). TiO2 film is directly deposited on one side of pure titanium disk as the

photoanode while Pt is deposited on the opposite side of the disk as the cathode. The titanium disk is mounted between

two Plexiglas compartments, under which a cation exchange membrane is placed allowing proton exchange between

different electrolytes on two sides. In another photocell built by Ida et al., the TiO2 photoanode and the CaFe2O4

photocathode are placed in two quartz cells (Figure 23b), which are interconnected by a Nafion 117 film(Ida et al., 2010).

Both quartz cells are illuminated by Xe lamp. Nafion is commonly used as the ion-permeable membrane but Minggu et

al.(Minggu et al., 2010) pointed out that Nafion or other cation exchange membranes is not proper for cation-containing

electrolytes as cations (eg. Na+) will replace H+ in the membrane and hinder the movement of protons through the

membrane. In such cases, other separators can be used to replace cation exchange membranes, which will be discussed in

the following.

Page 33 of 51  

Figure 23. Ion-permeable membrane separated PEC cells: (a) TiO2 photoanode and Pt cathode are deposited onto two sides of a titanium disk, which is mounted between two compartments(Selli et al., 2007) and (b)two quartz cells containing TiO2 photoanode and the CaFe2O4 photocathode respectively are interconnected by a Nafion 117 film(Ida et al., 2010). Note: in Figure 25a, different parts marked with numbers are 1 cation exchange membrane; 2 glass filter window; 3 burette; 4 stopcock; 5 rubber septum; and 6 reservior.

4.2.2.2. PEC cells with other separators  Besides ion-permeable membranes, other separators can be used to separate the hydrogen-evolving chamber and the

oxygen-evolving chamber (Figure 24), such as glass frit diaphragms and others(Kelly and Gibson, 2008; Li et al., 2011;

Murphy et al., 2006a). Kelly et al. established a spherical tank PEC reactor, using acrylic plastic to separate the anode and

cathode chamber, which possessed better transparency, strength and durability, as shown in Figure 24a(Kelly and Gibson,

2008). In a typical three-electrode PEC reactor, Murphy et al. exploited porous glass frit to isolate the compartment

containing counter electrode from the compartment containing the working electrode and reference electrode (Figure

24b)(Murphy et al., 2006b). Glass frit has the advantage of thermal, chemical and mechanical stability, and thus is a good

choice. As shown in Figure 24c, in another PEC cell configuration build from membrane electrode assembly (MEA),

anode (TiO2 anode) and cathode (Pt/C carbon paper) were placed on two sides of asbestos diaphragm(Li et al., 2011).

Asbestos diaphragm is widely used in alkaline electrolyzer due to its good hydrophilicity and ability to suppress back-

diffusion of OH-, but its application is limited because of its swelling properties and chemical instability under high

electric current.

Page 34 of 51  

Figure 24. PEC cells with other separators: a) spherical tank PEC reactor with acrylic plastic as chamber separator(Kelly and Gibson, 2008), b) three-electrode PEC cell with glass frit as separator(Murphy et al., 2006a) and c) membrane electrode assembly with asbestos diaphragm between anode and cathode(Li et al., 2011). Note: membrane in c) is asbestos diaphragm.

4.2.2.3. None­membrane separated PEC cells  When gas bubbles are formed on the surface of electrodes, they will always go up in the electrolyte. According to this

principle, some photocells are designed without the existence of ion-permeable membranes. For examples, the gas-

collecting device can be placed closely above the photoelectrodes(Mishra et al., 2003; Mishra et al., 2007), shown as

Figure 25. In this three-electrode photocell, two inverted burettes are fixed in the Perspex cell right above the TiO2

photoanode and Pt cathode so that H2 and O2 bubbles will be collected by the burettes immediately after they desorbed

from the electrode surface. With similar design idea, the chambers for each photoelectrode can be made into vertical

tubular shapes and interconnected by short horizontal tube, shown as Figure 26, and this type of photocells are often called

“H-type” photocells(Allam et al., 2008; Honda, 2004).

Page 35 of 51  

Figure 25. None-membrane separated PEC cells in which gas-collecting device are placed closely above the photoelectrodes(Mishra et al., 2003). Note: CE: counter electrode terminal, RE: reference electrode terminal, WE: working electrode terminal, 1: TiO2 working electrode, 2: Pt counter electrode fused in Pyrex glass, 3: saturated calomel electrode, 4; inverted burettes for gas collection, 5: quartz window, 6: electrolyte, 7: Perspex cell.

Page 36 of 51  

Figure 26. H-type photocells: (a)H-type photocell built by interconnecting tubular chambers used in Grimes Group(Allam et al., 2008) and (b) the H-type photocell constructed by Fujishima and Honda in 1970s(Honda, 2004).

4.2.3. PEC cells with quartz window  To fabricate a PEC cell made totally from quartz or Pyrex glass is obviously very expensive, so assembling a PEC cell

with cheaper material is desirable. One way to realize this is to build a cell with cheap materials but equip it with quartz

window where the photons can come into cell and hit the photoelectrode(Li et al., 2011; Mishra et al., 2003; Mishra et al.,

2007; Murphy et al., 2006a; Selli et al., 2007). The PEC cells can be fabricated with glass, Perspex, Teflon etc., and the

quartz window can be fixed by nuts and bolts with O-ring or gasket.

4.3. Large­scale trials  Even though the relatively low efficiency and stability issues for many PECs are still not completely solved, researchers

have set foot in trials of potential large-scale applications.

Page 37 of 51  

4.3.1. p/n PEC cells  Aroutiounian et al.(Aroutiounian et al., 2005) proposed a photocell configuration containing two photoactive electrodes,

that is, photoanode and photocathode, shown as Figure 27. A parabolic mirror is applied to concentrate sunlight and the

two curved quartz windows also help increase light intensity. The two photoelectrodes are placed side-by-side in two

separated chambers, connected by ion-exchange membrane. This configuration features low cost of the photoelectrodes,

but the parabolic mirror is very expensive and additional water cooling is needed.

Figure 27. p/n PEC cells(Aroutiounian et al., 2005). (1) PEC cell body; (2) quartz windows; (3) photoanode (or photocathode); (4) electrolyte; (5) ion-exchange membrane; (6) cylindrical light concentrator. Fan et al.(Fan et al., 2007) also proposed another p/n PEC cell module shown in Figure 28. In their PEC, the photocell

body is made from light transmissive materials such as glass or plexiglas while the photoelectrodes comprise polymer

electrolyte membrane, semiconductor layer on the membrane facing away from electrolyte and water-permeable materials.

During the reaction, water molecules will go into the three-phase region in the photoelectrodes, and thus the

photoelectrode will not be in contact with electrolyte directly, protecting the photoelectrode from corrosion. Moreover, a

mirror is placed behind the PEC cell perpendicular to the direction of light beams in order to reflect the light and utilize

photons more efficiently.

Page 38 of 51  

Figure 28. p/n PEC cell with photoelectrode-protection.(Fan et al., 2007)

4.3.2. PV­driven PEC cells  PEC cells which are completely driven by PV devices have been developed by some researchers(Deng and Xu, 2004,

2005; PI et al., 2006). In this type of PEC cells, PV devices are used to provide the potential for water splitting and

different gas production chambers are separated by ion-exchange membranes.

Figure 29 shows the schematic diagrams of two types of PV-driven PEC cells. In Figure 29a, the photoelectrode, which

actually means the PV device, along with the membrane, separate the PEC cell into two compartments(Deng and Xu,

2004). When the sunlight shines upon the PV device through the top glass panel, a voltage of around or above 1.6 eV is

generated to drive water electrolysis. The produced H2 and O2 are then stored in the two compartments. Another PV-

driven PEC cell proposed by Deng et al. is shown as Figure 29b(Deng and Xu, 2005). In this module, a PV device is

placed above the chamber containing electrolyte and the chamber is divided into a H2 evolution sub-chamber and an O2

evolution sub-chamber by an ion-exchange membrane. Two metal layers deposited with hydrogen evolution catalyst and

oxygen evolution catalyst, respectively, are placed in the two sun-chambers, connected to the PV with wire. After the PV

is illuminated by light from the top, H2 and O2 are produced in the two sub-chambers.

Page 39 of 51  

Figure 29. PV-driven PEC cell: a)(Deng and Xu, 2004; PI et al., 2006) and b) (Deng and Xu, 2005; PI et al., 2006).

4.3.3. PV­biased PEC cells  The PV/photoelectrode PEC cells have also been studied(Miller et al., 2003). As shown in Figure 30, the photoactive

semiconductor is connected with a PV, which is used to provide additional bias, in series, forming a hybrid

photoelectrode. Such a photoelectrode can be fabricated from a stainless steel foil with hydrogen evolution co-catalyst on

the back surface. On the front surface of stainless steel, the deposition order is solid-state multijunction, conductive

transparent oxide interface layer and photoactive semiconductor. The photoactive semiconductor with suitable band

position for oxygen evolution forms a photoelectrochemical junction when in contact with electrolyte. This design

provides the opportunity to stack series-connected PVs in a side-by-side manner to offer adequate bias, and to ease the

fabrication of this hybrid photoelectrode due to simplified geometry.

Page 40 of 51  

Figure 30. PV-biased PEC cell(Miller et al., 2003).

4.3.4. DSSC­biased PEC cell  Similar to PV-biased PEC cells, researchers also fabricated PEC cells containing DSSC to provide additional bias(Graetzel

and Augustynski, 2005; Minggu et al., 2010; Park and Bard, 2006). In such scheme, the photoanode need to be connected

to the counter-electrode of DSSC, which can provide holes. The connection can be realized by wire (Figure 31a) or by

sharing the substrate for photoanode and counter-electrode of DSSC (Figure 31b). In the cell shown as Figure 31a, two

photosystems are connected in series. A photoanode on the left connected to the counter-electrode of DSSC on the right by

wire will produce oxygen gas in the left chamber, and the catalytic cathode on the right connected to the photoanode of the

DSSC by wire will produce hydrogen gas in the right chamber. In the cell shown as Figure 31b, the photoanode on the

right connects to the counter-electrode of DSSC by sharing the substrate, and the cathode on the left connects to the

photoanode of DSSC in the same way.

Page 41 of 51  

Figure 31. DSSC-biased PEC cells: a)(Graetzel and Augustynski, 2005) and b)(Minggu et al., 2010; Park and Bard, 2006).

4.4. Standardization of photoelectrochemical water splitting evaluation 

Similar to photocatalytic water splitting photoreactors, there are a few different configurations for PEC cells or reactors,

which makes it difficult to cross-compare the efficiency of materials. As research advances in this area expanded rapidly

recently, standardized evaluation methods is necessary to help screening of efficient materials and scientific

communication(Chen et al., 2010b). The three-electrode photocell has proven itself to be the most widely exploited

configuration in lab-scale of photoelectrode investigations due to its simplicity in set-up and measurement implementation.

As illustrated in part 4.2.1, Maeda’s work provided a very good example of three-electrode configuration for n-type

photoanode PEC and p-type photocathode PEC(Maeda et al., 2011). For this type of photocell set-up, several suggestions

are listed here:

Page 42 of 51  

(1) Light source: same to photochemical water splitting reactor, a solar simulator equipped with AM1.5 filter is the

best choice of light source.

(2) Photocell build-up: the side of photocell facing the lamp should be fabricated from materials with good light

transmission properties. Quartz is the mostly preferred. Different ports with fixed position should be built which

can accommodate different electrodes and make the measurement reproducible.

(3) Electrolyte: environment surrounding the photoelectrodes pose considerable impact on their

photoelectrochemical performance, so the composition of the electrolyte plays a vital role in PEC evaluation. As

a result, the condition of electrolyte must be clearly described along with the measured results.

For large-scale PEC water splitting cells, we also list two factors to consider:

(1) Illumination: Similar to large-scale photochemical reactor, the illumination condition under direct sunlight can be

very tricky, so the light intensity should be accurately reported.

(2) Light concentrating: When the area of photoactive electrode is fixed, the effective way to enhance the overall

performance is to increase the light intensity. Placing a light concentrating device in front of the glass window

panel is a very good idea shared with CPC photochemical reactors, which should be paid more attention to when

developing PEC cells.

5. Summary and outlook The past few decades witnessed the quick advancement of photoelectrochemical and photochemical water decomposition

for solar hydrogen production. From the materials aspect, the microscopic structure, morphology, crystal size, energy band

structure and crystal facet orientation of semiconductors can be well tuned via many advanced nanotechnologies and

highly controllable chemical methodologies. On the other hand, the engineering part of solar hydrogen production, that is,

photoreactor design and configuration, has yet been drawn enough attention to, especially the semiconductor-particles-

based photochemical water splitting, even though it greatly affects the overall hydrogen production. This paper

summarizes some typical photoreactor configurations for photochemical and photoelectrochemical water splitting used

both in lab-scale investigations and potential scale-up practices.

Page 43 of 51  

For photochemical water splitting, batch-type photoreactor is most frequently used configuration in lab-scale

investigations, which often consists of reaction tank, cooling system, evacuation module, gas detection instrument etc.

Each part of a batch-type photoreactor can influence the performance in practical operation, which causes great difficulty

in standardizing photocatalyst evaluation. So far, there is no widely recognized standard testing photoreactors. For lab-

scale study, the most important thing is to accurately control as many factors in the reaction system as possible. Against

such background, a masked cylindrical batch type reactor equipped with fixed quartz window is suggested because of its

easy fabrication, relatively low cost and well controlled geometry. Proper temperature control system and evacuation

system must be applied, while a solar simulator equipped with an AM 1.5 filter should be used as the light source, whose

light intensity is measured by actinometry. In spite of the still low efficiency of direct photocatalytic water splitting, some

researchers have tried to modify photoreactor set-up in order to improve hydrogen production rate, but a lot more work is

in need. Few trials in practical or large-scale application can be found. Different from lab-scale study, the ultimate goal of

practical trials is to maximize the overall system hydrogen evolution yield. The CPC photoreactor provides an excellent

example of large-scale water photolysis reactor owing to its ability of concentrating sunlight.

For photoelectrochemical water splitting, three-electrode PEC set-up has shown its prevalence in lab-scale research on p-

type or n-type semiconductor photoelectrode PECs to determine solar-to-hydrogen efficiency. Other different geometries

of photocell vessel have also been developed to obtain higher energy conversion efficiency through maximizing light

capture and effectively separating products while reducing the fabrication cost at the same time. The author believes that

three-electrode PEC remains the best configuration, and the photocell with fixed ports and quartz window promises an

effective way towards the standardization. Some modules are already patented in large-scale application, though more

investigation is in demand. In spite of the relative high fabrication cost, the p/n PEC cell shown in Figure 27 seems to be

promising considering its photon collecting efficiency.

For both photochemical and photoelectrochemical water splitting, discovery of more visible-light-active semiconductor

materials with suitable energy band structure for water splitting and further unveiling of photocatalytic reaction

mechanism will contribute to potential large-scale applications. Standardization of the evaluation method is urgently

needed to identify promising photoactive materials. Lab-scale investigation and potential industrial application have

different requirements from the photoreactor design, as the former needs as high accuracy and repeatability as possible

while the latter demands superior photon absorption and gaseous product collection. Many precedent photoreactors for

water purification provide good examples of light harvesting for water splitting, so proper adjustment of water treatment

Page 44 of 51  

photoreactors may promise a feasible way in approaching large scale utility. So far there are already various PEC

photocell configurations for practical applications, yet a simpler design with maximum exposure to light and low

fabrication cost is desired.

The area of solar hydrogen production has attracted interest from all around the world, more research and investigation

will be focused on this area in the future. With further development of new photocatalyst, photoelectrode fabrication

technology and photoreactor optimization, the solar hydrogen production has the potential to solve many energy-related

issues.

Acknowledgements  The financial support from Australian Research Council (ARC) DP and LP programs is highly appreciated.   

References   Alfano, O., Bahnemann, D., Cassano, A., Dillert, R., Goslich, R., 2000. Photocatalysis in water environments using artificial and solar light. Catalysis today 58, 199‐230. Alic, J., Sarewitz, D., Weiss, C., Bonvillian, W., 2010. A new strategy for energy innovation. Nature 466, 316‐317. Allam, N.K., Shankar, K., Grimes, C.A., 2008. Photoelectrochemical and water photoelectrolysis properties of ordered TiO2 nanotubes fabricated by Ti anodization in fluoride‐free HCl electrolytes. Journal of Materials Chemistry 18, 2341‐2348. Aroutiounian, V., Arakelyan, V., Shahnazaryan, G., 2005. Metal oxide photoelectrodes for hydrogen generation using solar radiation‐driven water splitting. Solar Energy 78, 581‐592. Bahnemann, D., 2004. Photocatalytic water treatment: solar energy applications. Solar Energy 77, 445‐459. Bahnemann, D.W., 2000. Current challenges in photocatalysis: Improved photocatalysts and appropriate photoreactor engineering. Research on Chemical Intermediates 26, 207‐220. Bai, Y., Yu, H., Li, Z., Amal, R., Lu, G.Q., Wang, L., 2012. In Situ Growth of a ZnO Nanowire Network within a TiO2 Nanoparticle Film for Enhanced Dye‐Sensitized Solar Cell Performance. Advanced Materials 24, 5850‐5856. Braham, R.J., Harris, A.T., 2009. Review of major design and scale‐up considerations for solar photocatalytic reactors. Industrial & Engineering Chemistry Research 48, 8890‐8905. 

Page 45 of 51  

Brandi, R., Alfano, O., Cassano, A., 2000. Evaluation of radiation absorption in slurry photocatalytic reactors. 2. Experimental verification of the proposed method. Environmental science & technology 34, 2631‐2639. Brennan, P., Fedor, C., Pausch, G., 1988. Sunlight, UV and accelerated weathering. Paint and Resin 58, 17. Brown, K.A., Wilker, M.B., Boehm, M., Dukovic, G., King, P.W., 2012. Characterization of Photochemical Processes for H2 Production by CdS Nanorod–[FeFe] Hydrogenase Complexes. Journal of the American Chemical Society 134, 5627‐5636. Carver, C., Ulissi, Z., Ong, C., Dennison, S., Kelsall, G., Hellgardt, K., 2012. Modelling and development of photoelectrochemical reactor for H2 production. International Journal of Hydrogen Energy 37, 2911‐2923. Chen, J.J., Wu, J.C.S., Wu, P.C., Tsai, D.P., 2011. Plasmonic Photocatalyst for H2 Evolution in Photocatalytic Water Splitting. Journal of Physical Chemistry C 115, 210‐216. Chen, X., Shen, S., Guo, L., Mao, S.S., 2010a. Semiconductor‐based photocatalytic hydrogen generation. Chemical reviews 110, 6503‐6570. Chen, Z., Jaramillo, T.F., Deutsch, T.G., Kleiman‐Shwarsctein, A., Forman, A.J., Gaillard, N., Garland, R., Takanabe, K., Heske, C., Sunkara, M., 2010b. Accelerating materials development for photoelectrochemical hydrogen production: Standards for methods, definitions, and reporting protocols. J. Mater. Res 25, 3‐16. Deng, X., Xu, L., 2004. Integrated Photoelectrochemical cell and system having a liquid electrolyte. WO Patent WO/2004/050,961. Deng, X., Xu, L., 2005. Interconnected Photoelectrochemical Cell. WO Patent WO/2005/081,326. Dresselhaus, M.S., Thomas, I.L., 2001. Alternative energy technologies. Nature 414, 332‐337. Escudero, J., Cervera‐March, S., Giménez, J., Simarro, R., 1990. Preparation and characterization of Pt(RuO2)/TiO2 catalysts: Test in a continuous water photolysis system. Journal of Catalysis 123, 319‐332. Escudero, J., Gimenez, J., Simarro, R., Cervera‐March, S., 1988. Physical characteristics of photocatalysts affecting the performance of a process in a continuous photoreactor. Solar energy materials 17, 151‐163. Escudero, J., Simarro, R., Cervera‐March, S., Giménez, J., 1989. Rate‐controlling steps in a three‐phase (solid—liquid—gas) photoreactor: a phenomenological approach applied to hydrogen photoprodution using Pt‐TiO2 aqueous suspensions. Chemical Engineering Science 44, 583‐593. Esplugas, S., Cervera, S., Simarro, R., 1987. A reactor model for water photolysis experimental studies in the liquid phase with suspensions of catalytic particles. Chemical engineering communications 51, 221‐232. Fan, Q., Onischak, M., Liss, W.E., 2007. Solar cell electrolysis of water to make hydrogen and oxygen. Google Patents. Fujishima, A., Honda, K., 1972. Electrochemical photolysis of water at a semiconductor electrode. Nature 238, 37‐38. Graetzel, M., Augustynski, J., 2005. Tandem cell for water cleavage by visible light. Google Patents. Grätzel, M., 2001. Photoelectrochemical cells. 

Page 46 of 51  

Guo, L.J., Zhao, L., Jing, D.W., Lu, Y.J., Yang, H.H., Bai, B.F., Zhang, X.M., Ma, L.J., Wu, X.M., 2010. Reprint of: Solar hydrogen production and its development in China. Energy 35, 4421‐4438. Hagfeldt, A., Boschloo, G., Sun, L.C., Kloo, L., Pettersson, H., 2010. Dye‐Sensitized Solar Cells. Chemical reviews 110, 6595‐6663. Han, Z., Qiu, F., Eisenberg, R., Holland, P.L., Krauss, T.D., 2012. Robust Photogeneration of H2 in Water Using Semiconductor Nanocrystals and a Nickel Catalyst. Science 338, 1321‐1324. Hara, M., Nunoshige, J., Takata, T., Kondo, J.N., Domen, K., 2003. Unusual enhancement of H‐2 evolution by Ru on TaON photocatalyst under visible light irradiation. Chemical Communications, 3000‐3001. Herrmann, J.‐M., 2005. Heterogeneous photocatalysis: state of the art and present applications. Topics in Catalysis 34, 49‐65. Higashi, M., Domen, K., Abe, R., 2012. Highly Stable Water Splitting on Oxynitride TaON Photoanode System under Visible Light Irradiation. Journal of the American Chemical Society 134, 6968‐6971. Hisatomi, T., Maeda, K., Takanabe, K., Kubota, J., Domen, K., 2009. Aspects of the Water Splitting Mechanism on (Ga1−xZnx)(N1−xO x) Photocatalyst Modified with Rh2−yCryO3 Cocatalyst. The Journal of Physical Chemistry C 113, 21458‐21466. Hoffmann, M.R., Martin, S.T., Choi, W.Y., Bahnemann, D.W., 1995. Environmental applications of semiconductor photocatalysis. Chemical reviews 95, 69‐96. Honda, K., 2004. Dawn of the evolution of photoelectrochemistry. Journal of Photochemistry and Photobiology A: Chemistry 166, 63‐68. Hou, Y., Abrams, B.L., Vesborg, P.C.K., Björketun, M.E., Herbst, K., Bech, L., Setti, A.M., Damsgaard, C.D., Pedersen, T., Hansen, O., 2011. Bioinspired molecular co‐catalysts bonded to a silicon photocathode for solar hydrogen evolution. Nature materials 10, 434‐438. <http://assets.newport.com/webDocuments‐EN/images/12298.pdf>. <http://www.eyesolarlux.com/Solar‐simulation‐ASTM‐IEC‐JIS.htm>. Manke, E., 2011. Standards For Simulators Can Vary Widely<http://www.newport.com/images/webdocuments‐en/images/Solar_Industry‐Solar_Simulation.pdf>.  <http://www.newport.com/Oriel‐Sol3A‐Class‐AAA‐Solar‐Simulators/842468/1033/info.aspx#tab_Overview>. <http://www.yamashitadenso.co.jp/english/product03.html>. Huang, C., Yao, W., 2011. Development of efficient photoreactors for solar hydrogen production. Solar Energy 85, 19‐27. Huang, Q.S., Liu, T.Z., Yang, J., Yao, L.S., Gao, L.L., 2011. Evaluation of radiative transfer using the finite volume method in cylindrical photoreactors. Chemical Engineering Science 66, 3930‐3940. Ida, S., Yamada, K., Matsunaga, T., Hagiwara, H., Matsumoto, Y., Ishihara, T., 2010. Preparation of p‐Type CaFe2O4 Photocathodes for Producing Hydrogen from Water. Journal of the American Chemical Society 132, 17343‐17345. Ishikawa, A., Takata, T., Kondo, J.N., Hara, M., Kobayashi, H., Domen, K., 2002. Oxysulfide Sm2Ti2S2O5 as a stable photocatalyst for water oxidation and reduction under visible light irradiation (λ≤ 650 nm). Journal of the American Chemical Society 124, 13547‐13553. 

Page 47 of 51  

Ishikawa, A., Takata, T., Matsumura, T., Kondo, J.N., Hara, M., Kobayashi, H., Domen, K., 2004. Oxysulfides Ln2Ti2S2O5 as stable photocatalysts for water oxidation and reduction under visible‐light irradiation. The Journal of Physical Chemistry B 108, 2637‐2642. James, B.D., Baum, G.N., Perez, J., Baum, K.N., Square, O.V., 2009. Technoeconomic Analysis of Photoelectrochemical (PEC) Hydrogen Production. DOE Report. Jing, D., Guo, L., Zhao, L., Zhang, X., Liu, H., Li, M., Shen, S., Liu, G., Hu, X., 2010. Efficient solar hydrogen production by photocatalytic water splitting: From fundamental study to pilot demonstration. International Journal of Hydrogen Energy 35, 7087‐7097. Jing, D., Liu, H., Zhang, X., Zhao, L., Guo, L., 2009. Photocatalytic hydrogen production under direct solar light in a CPC based solar reactor: Reactor design and preliminary results. Energy Conversion and Management 50, 2919‐2926. Kamat, P.V., 2007. Meeting the clean energy demand: Nanostructure architectures for solar energy conversion. Journal of Physical Chemistry C 111, 2834‐2860. Karuturi, S.K., Luo, J., Cheng, C., Liu, L., Su, L.T., Tok, A.I.Y., Fan, H.J., 2012a. A Novel Photoanode with Three‚ÄêDimensionally, Hierarchically Ordered Nanobushes for Highly Efficient Photoelectrochemical Cells. Advanced Materials. Karuturi, S.K., Luo, J.S., Cheng, C.W., Liu, L.J., Su, L.T., Tok, A.I.Y., Fan, H.J., 2012b. A Novel Photoanode with Three‐Dimensionally, Hierarchically Ordered Nanobushes for Highly Efficient Photoelectrochemical Cells. Advanced Materials 24, 4157‐4162. Kato, H., Asakura, K., Kudo, A., 2003. Highly efficient water splitting into H2 and O2 over lanthanum‐doped NaTaO3 photocatalysts with high crystallinity and surface nanostructure. Journal of the American Chemical Society 125, 3082‐3089. Kelly, N.A., Gibson, T.L., 2006. Design and characterization of a robust photoelectrochemical device to generate hydrogen using solar water splitting. International Journal of Hydrogen Energy 31, 1658‐1673. Kelly, N.A., Gibson, T.L., 2008. Solar energy concentrating reactors for hydrogen production by photoelectrochemical water splitting. International Journal of Hydrogen Energy 33, 6420‐6431. Khnayzer, R.S., Thompson, L.B., Zamkov, M., Ardo, S., Meyer, G.J., Murphy, C.J., Castellano, F.N., 2011. Photocatalytic Hydrogen Production at Titania‐Supported Pt Nanoclusters That Are Derived from Surface‐Anchored Molecular Precursors. The Journal of Physical Chemistry C 116, 1429‐1438. Khnayzer, R.S., Thompson, L.B., Zamkov, M., Ardo, S., Meyer, G.J., Murphy, C.J., Castellano, F.N., 2012. Photocatalytic Hydrogen Production at Titania‐Supported Pt Nanoclusters That Are Derived from Surface‐Anchored Molecular Precursors. Journal of Physical Chemistry C 116, 1429‐1438. Kim, H.G., Hwang, D.W., Lee, J.S., 2004. An undoped, single‐phase oxide photocatalyst working under visible light. Journal of the American Chemical Society 126, 8912‐8913. Kim, J., Kim, H.R., Choung, S.J., 2011. The Characterization and Hydrogen Production from Water Decomposition with Methanol in a Semi‐Batch Type Reactor Using In, P‐TiO 2 s. International Journal of Photoenergy 2011. Kudo, A., Miseki, Y., 2008. Heterogeneous photocatalyst materials for water splitting. Chem. Soc. Rev. 38, 253‐278. Kuhn, H.J., Braslavsky, S.E., Schmidt, R., 2004. Chemical actinometry. Pure and Applied Chemistry 76, 2105‐2146. 

Page 48 of 51  

Lewis, N.S., 2005. Chemical control of charge transfer and recombination at semiconductor photoelectrode surfaces. Inorganic Chemistry 44, 6900‐6911. Lewis, N.S., Nocera, D.G., 2006. Powering the planet: Chemical challenges in solar energy utilization. Proceedings of the National Academy of Sciences of the United States of America 103, 15729‐15735. Li, Y., Yu, H., Song, W., Li, G., Yi, B., Shao, Z., 2011. A novel photoelectrochemical cell with self‐organized TiO2 nanotubes as photoanodes for hydrogen generation. International Journal of Hydrogen Energy 36, 14374‐14380. Liu, G., Niu, P., Sun, C.H., Smith, S.C., Chen, Z.G., Lu, G.Q., Cheng, H.M., 2010a. Unique Electronic Structure Induced High Photoreactivity of Sulfur‐Doped Graphitic C3N4. Journal of the American Chemical Society 132, 11642‐11648. Liu, G., Wang, L., Yang, H.G., Cheng, H.M., Lu, G.Q.M., 2009. Titania‐based photocatalysts—crystal growth, doping and heterostructuring. Journal of Materials Chemistry 20, 831‐843. Liu, G., Wang, L., Yang, H.G., Cheng, H.M., Lu, G.Q.M., 2010b. Titania‐based photocatalysts—crystal growth, doping and heterostructuring. Journal of Materials Chemistry 20, 831‐843. Lo, C.C., Huang, C.W., Liao, C.H., Wu, J., 2010. Novel twin reactor for separate evolution of hydrogen and oxygen in photocatalytic water splitting. International Journal of Hydrogen Energy 35, 1523‐1529. Luo, J., Ma, L., He, T., Ng, C.F., Wang, S., Sun, H., Fan, H.J., 2012. TiO2/(CdS, CdSe, CdSeS) Nanorod Heterostructures and Photoelectrochemical Properties. The Journal of Physical Chemistry C 116, 11956‐11963. Maeda, K., 2011. Photocatalytic water splitting using semiconductor particles: History and recent developments. Journal of Photochemistry and Photobiology C‐Photochemistry Reviews 12, 237‐268. Maeda, K., Domen, K., 2011. Surface Nanostructures in Photocatalysts for Visible‐Light‐Driven Water Splitting, in: Bignozzi, C.A. (Ed.), Photocatalysis. Springer‐Verlag Berlin, Berlin, pp. 95‐119. Maeda, K., Higashi, M., Siritanaratkul, B., Abe, R., Domen, K., 2011. SrNbO2N as a Water‐Splitting Photoanode with a Wide Visible‐Light Absorption Band. Journal of the American Chemical Society 133, 12334‐12337. Maeda, K., Teramura, K., Domen, K., 2008. Effect of post‐calcination on photocatalytic activity of (Ga1‐xZnx)(N1‐xOx) solid solution for overall water splitting under visible light. Journal of Catalysis 254, 198‐204. Maeda, K., Teramura, K., Lu, D.L., Takata, T., Saito, N., Inoue, Y., Domen, K., 2006a. Characterization of Rh‐Cr mixed‐oxide nanoparticles dispersed on (Ga1‐xZnx)(N1‐xOx) as a cocatalyst for visible‐light‐driven overall water splitting. Journal of Physical Chemistry B 110, 13753‐13758. Maeda, K., Teramura, K., Lu, D.L., Takata, T., Saito, N., Inoue, Y., Domen, K., 2006b. Photocatalyst releasing hydrogen from water ‐ Enhancing catalytic performance holds promise for hydrogen production by water splitting in sunlight. Nature 440, 295‐295. Maeda, K., Teramura, K., Masuda, H., Takata, T., Saito, N., Inoue, Y., Domen, K., 2006c. Efficient overall water splitting under visible‐light irradiation on (Ga1‐xZnx)( N1‐xOx) dispersed with Rh‐Cr mixed‐oxide nanoparticles: Effect of reaction conditions on photocatalytic activity. Journal of Physical Chemistry B 110, 13107‐13112. 

Page 49 of 51  

Mangrulkar, P.A., Joshi, M.V., Kamble, S.P., Labhsetwar, N.K., Rayalu, S.S., 2010. Hydrogen evolution by a low cost photocatalyst: Bauxite residue. International Journal of Hydrogen Energy 35, 10859‐10866. Mangrulkar, P.A., Polshettiwar, V., Rayalu, S.S., 2012. Nano Ferrites for Water Splitting: Unprecedentedly High Photocatalytic Hydrogen Production under Visible Light. Nanoscale. Marschall, R., Klaysom, C., Mukherji, A., Wark, M., Lu, G.Q.M., Wang, L., 2011a. Composite proton‐conducting polymer membranes for clean hydrogen production with solar light in a simple photoelectrochemical compartment cell. International Journal of Hydrogen Energy. Marschall, R., Mukherji, A., Tanksale, A., Sun, C., Smith, S.C., Wang, L., Lu, G.Q.M., 2011b. Preparation of new sulfur‐doped and sulfur/nitrogen co‐doped CsTaWO6 photocatalysts for hydrogen production from water under visible light. Journal of Materials Chemistry 21, 8871‐8879. Maschmeyer, T., Che, M., 2010. Catalytic Aspects of Light‐Induced Hydrogen Generation in Water with TiO2 and Other Photocatalysts: A Simple and Practical Way Towards a Normalization? Angewandte Chemie International Edition 49, 1536‐1539. McCullagh, C., Skillen, N., Adams, M., Robertson, P.K.J., 2011. Photocatalytic reactors for environmental remediation: a review. Journal of Chemical Technology and Biotechnology 86, 1002‐1017. Miller, E., Rocheleau, R., 1999. Photoelectrochemical hydrogen production. Proceedings of the 1999 DOE Hydrogen program review, NREL/CP‐570‐26938. Miller, E., Rocheleau, R., Deng, X., 2003. Design considerations for a hybrid amorphous silicon/photoelectrochemical multijunction cell for hydrogen production. International Journal of Hydrogen Energy 28, 615‐623. Mills, A., LeHunte, S., 1997. An overview of semiconductor photocatalysis. Journal of Photochemistry and Photobiology a‐Chemistry 108, 1‐35. Minggu, L.J., Wan Daud, W.R., Kassim, M.B., 2010. An overview of photocells and photoreactors for photoelectrochemical water splitting. International Journal of Hydrogen Energy 35, 5233‐5244. Mishra, P., Shukla, P., Singh, A., Srivastava, O., 2003. Investigation and optimization of nanostructured TiO2 photoelectrode in regard to hydrogen production through photoelectrochemical process. International Journal of Hydrogen Energy 28, 1089‐1094. Mishra, P., Shukla, P., Srivastava, O., 2007. Study of modular PEC solar cells for photoelectrochemical splitting of water employing nanostructured TiO2 photoelectrodes. International Journal of Hydrogen Energy 32, 1680‐1685. Morton, O., 2006. Solar energy: A new day dawning?: Silicon Valley sunrise. Nature 443, 19‐22. Mukherji, A., Marschall, R., Tanksale, A., Sun, C., Smith, S.C., Lu, G.Q., Wang, L., 2011a. N‐Doped CsTaWO6 as a New Photocatalyst for Hydrogen Production from Water Splitting Under Solar Irradiation. Advanced Functional Materials 21, 126‐132. Mukherji, A., Marschall, R., Tanksale, A., Sun, C.H., Smith, S.C., Lu, G.Q., Wang, L.Z., 2011b. N‐Doped CsTaWO6 as a New Photocatalyst for Hydrogen Production from Water Splitting Under Solar Irradiation. Advanced Functional Materials 21, 126‐132. Murphy, A., Barnes, P., Randeniya, L., Plumb, I., Grey, I., Horne, M., Glasscock, J., 2006a. Efficiency of solar water splitting using semiconductor electrodes. International Journal of Hydrogen Energy 31, 1999‐2017. 

Page 50 of 51  

Murphy, A.B., Barnes, P.R.F., Randeniya, L.K., Plumb, I.C., Grey, I.E., Horne, M.D., Glasscock, J.A., 2006b. Efficiency of solar water splitting using semiconductor electrodes. International Journal of Hydrogen Energy 31, 1999‐2017. Nowotny, J., Sorrell, C.C., Bak, T., Sheppard, L.R., 2005. Solar‐hydrogen: Unresolved problems in solid‐state science. Solar Energy 78, 593‐602. Nozik, A.J., 1977. PHOTOCHEMICAL DIODES. Applied Physics Letters 30, 567‐569. Oralli, E., Dincer, I., Naterer, G., 2011. Solar photocatalytic reactor performance for hydrogen production from incident ultraviolet radiation. International Journal of Hydrogen Energy 36, 9446‐9452. Pan, J., Liu, G., Lu, G.M., Cheng, H.M., 2011a. On the True Photoreactivity Order of {001}, {010}, and {101} Facets of Anatase TiO2 Crystals. Angewandte Chemie‐International Edition 50, 2133‐2137. Pan, J., Wu, X., Wang, L.Z., Liu, G., Lub, G.Q., Cheng, H.M., 2011b. Synthesis of anatase TiO2 rods with dominant reactive {010} facets for the photoreduction of CO2 to CH4 and use in dye‐sensitized solar cells. Chemical Communications 47, 8361‐8363. Paracchino, A., Laporte, V., Sivula, K., Grätzel, M., Thimsen, E., 2011. Highly active oxide photocathode for photoelectrochemical water reduction. Nature materials 10, 456‐461. Park, J.H., Bard, A.J., 2006. Photoelectrochemical tandem cell with bipolar dye‐sensitized electrodes for vectorial electron transfer for water splitting. Electrochemical and solid‐state letters 9, E5‐E8. Park, J.H., Kim, S., Bard, A.J., 2006. Novel carbon‐doped TiO2 nanotube arrays with high aspect ratios for efficient solar water splitting. Nano letters 6, 24‐28. Parker, C.A., 1953. A New Sensitive Chemical Actinometer. I. Some Trials with Potassium Ferrioxalate. Proceedings of the Royal Society of London Series a‐Mathematical and Physical Sciences 220, 104‐116. PI, L.X., Midwest Optoelectronics, L., Ingler Jr, W.B., Abken, A., 2006. Critical Research for Cost‐Effective Photoelectrochemical Production of Hydrogen. Priya, R., Kanmani, S., 2009. Batch slurry photocatalytic reactors for the generation of hydrogen from sulfide and sulfite waste streams under solar irradiation. Solar Energy 83, 1802‐1805. Quan, Y., Pehkonen, S.O., Madhumita, B., 2004. Evaluation of three different lamp emission models using novel application of potassium ferrioxalate actinometry. Industrial & Engineering Chemistry Research 43, 948‐955. Reber, J.F., Meier, K., 1984. Photochemical production of hydrogen with zinc‐sulfide suspensions. Journal of Physical Chemistry 88, 5903‐5913. Romero, R.L., Alfano, O.M., Cassano, A.E., 1997. Cylindrical photocatalytic reactors. Radiation absorption and scattering effects produced by suspended fine particles in an annular space. Industrial & Engineering Chemistry Research 36, 3094‐3109. Sabate, J., Cervera‐March, S., Simarro, R., Gimenez, J., 1990. Photocatalytic production of hydrogen from sulfide and sulfite waste streams: a kinetic model for reactions occurring in illuminating suspensions of CdS. Chemical Engineering Science 45, 3089‐3096. Sayama, K., Arakawa, H., 1997. Effect of carbonate salt addition on the photocatalytic decomposition of liquid water over Pt–TiO2 catalyst. Journal of the Chemical Society, Faraday Transactions 93, 1647‐1654. 

Page 51 of 51  

Seger, B., Lu, G.Q.M., Wang, L., 2012. Electrical power and hydrogen production from a photo‐fuel cell using formic acid and other single‐carbon organics. Journal of Materials Chemistry. Selli, E., Chiarello, G.L., Quartarone, E., Mustarelli, P., Rossetti, I., Forni, L., 2007. A photocatalytic water splitting device for separate hydrogen and oxygen evolution. Chemical Communications, 5022‐5024. Tachibana, Y., Vayssieres, L., Durrant, J.R., 2012. Artificial photosynthesis for solar water‐splitting. Nature Photonics 6, 511‐518. Tetreault, N., Arsenault, E., Heiniger, L.P., Soheilnia, N., Brillet, J., Moehl, T., Zakeeruddin, S., Ozin, G.A., Gratzel, M., 2011. High‐Efficiency Dye‐Sensitized Solar Cell with Three‐Dimensional Photoanode. Nano letters 11, 4579‐4584. Tseng, C.‐J., Wang, C.‐H., Cheng, K.‐W., 2011. Photoelectrochemical performance of gallium‐doped AgInS2 photoelectrodes prepared by electrodeposition process. Solar Energy Materials and Solar Cells. Turner, J.A., 1999. A realizable renewable energy future. Science 285, 687‐689. Varner, K., Warren, S., Turner, J., 2002. Photoelectrochemical systems for hydrogen production. Vidyarthi, V.S., Hofmann, M., Savan, A., Sliozberg, K., König, D., Beranek, R., Schuhmann, W., Ludwig, A., 2011. Enhanced photoelectrochemical properties of WO3 thin films fabricated by reactive magnetron sputtering. International Journal of Hydrogen Energy 36, 4724‐4731. Walter, M.G., Warren, E.L., McKone, J.R., Boettcher, S.W., Mi, Q., Santori, E.A., Lewis, N.S., 2010. Solar water splitting cells. Chemical reviews 110, 6446. Wang, X., Liu, G., Chen, Z.G., Li, F., Lu, G.Q., Cheng, H.M., 2009. Efficient and stable photocatalytic H2 evolution from water splitting by Cd0.8Zn0.2S nanorods. Electrochemistry Communications 11, 1174‐1178. Wu, X., Chen, Z., Lu, G.Q.M., Wang, L., 2011a. Nanosized Anatase TiO2 Single Crystals with Tunable Exposed (001) Facets for Enhanced Energy Conversion Efficiency of Dye‐Sensitized Solar Cells. Advanced Functional Materials 21, 4167‐4172. Wu, X., Lu, G.Q.M., Wang, L., 2011b. Shell‐in‐shell TiO2 hollow spheres synthesized by one‐pot hydrothermal method for dye‐sensitized solar cell application. Energy & Environmental Science 4, 3565‐3572. Yang, Q., Ling Ang, P., Ray, M.B., Pehkonen, S.O., 2005. Light distribution field in catalyst suspensions within an annular photoreactor. Chemical Engineering Science 60, 5255‐5268. Yu, S.C., Huang, C.W., Liao, C.H., Wu, J., Chang, S.T., Chen, K.H., 2011. A novel membrane reactor for separating hydrogen and oxygen in photocatalytic water splitting. Journal of Membrane Science 382, 291‐299. Zhou, H., Sabio, E.M., Townsend, T.K., Fan, T., Zhang, D., Osterloh, F.E., 2010. Assembly of Core− Shell Structures for Photocatalytic Hydrogen Evolution from Aqueous Methanol. Chemistry of Materials 22, 3362‐3368. Zong, X., Xing, Z., Yu, H., Chen, Z., Tang, F., Zou, J., Lu, G.Q., Wang, L., 2011. Photocatalytic water oxidation on F, N co‐doped TiO2 with dominant exposed {001} facets under visible light. Chemical Communications.