Photoelectrochemical Cell Design, Efficiency, Definitions, Standards, and Protocols

Chapter 4
Photoelectrochemical Cell Design, Efficiency,

Definitions, Standards, and Protocols
Wilson A. Smith
4.1
Introduction
The storage of solar energy into chemical energy through photoelectrochemical

water splitting offers a long-term, sustainable, and effective solution to the global
energy and environmental problems (Lewis and Nocera 2006). It has been over
40 years since the discovery of electrochemical photolysis of water (Fujishima and
Honda 1972), and yet today no commercial or industrial device exists that is
effectively producing solar hydrogen. The major limitations of the technological
advancement of this field are due to the complicated physical, chemical, and
engineering feats that are required to convert photons into electrons that can
directly drive electrochemical reactions with well-defined, separated, and transportable products. While a commercial solar fuel device is not presently contributing to the global energy supply, many attempts have been made to understand the
mechanisms and limitations of the photo-physical and chemical problems, and
many different arrangements of lab-scale devices have been explored.
The two biggest hurdles to accomplish the development of a practical PEC
water-splitting device are both scientific and technical. From a scientific standpoint,
a PEC device needs to be able to manage solar irradiation, transport electric and
ionic charges, and perform oxidative and reductive catalytic reactions simultaneously. From a technological point of view, these challenges must be all addressed
using materials and fabrication processes which are cheap and scalable, putting
severe limitations on the methods and compounds that can be used. However, since
the scientific challenges to achieve stable, efficient, and cost-effective PEC water
splitting have still not been overcome to a sufficient level, it may be pre-emptive to
begin the design of a practical system larger than a lab-scale device. Therefore,
W.A. Smith (*)
Materials for Energy Conversion and Storage (MECS), Department of Chemical Engineering,
Faculty of Applied Sciences, Delft University of Technology, 2628 BL Delft, The Netherlands
e-mail: [email protected]
Springer International Publishing Switzerland 2016
S. Gimenez, J. Bisquert (eds.), Photoelectrochemical Solar Fuel Production,
DOI 10.1007/978-3-319-29641-8_4
163
164
W.A. Smith
there is still a need to focus on the basic understanding, system and materials
diagnostics, and fundamental mechanisms involved in PEC water splitting.
For the purposes of this chapter, only the PEC water-splitting reaction will be
discussed and will not be compared to a solar fuel device that also carries out CO2
reduction (where the corresponding oxidation reaction is still water oxidation). The
overall water-splitting reaction can be summarized in the following equation:
2H2 O l ! 2H2 g O2 g
Eo 1:23 V G 237:22 kJ=mol

4:1
which shows that, theoretically, it takes a minimum of 1.23 V to split water into
molecular hydrogen and oxygen at the standard temperature (T 0 298 K) and
pressure (P0 1 bar). In practice, it takes several hundreds of mV overpotential
to drive the water-splitting reaction, mainly due to overpotentials associated with
water oxidation (Rossmeisl et al. 2007; Koper 2011), but can also depend on the
electrode material(s) used, the electrolyte, the distance between the electrodes, and
the device geometry.
The overall reaction takes place simultaneously at two different sites, which
mediate the oxidation reaction at an anode and the reduction reaction at a cathode.
In an acidic environment (pH 0), the two relevant half-reactions can be written:
2H2 O l ! 4H aq 4e O2 g Eo 1:23 V vs: NHE
4H aq 4 e ! 2H2 g
E 0:00 V vs: NHE

o
4:2
4:3
In an alkaline environment (pH 14), these red-ox equations then become:

4OH aq ! 2H2 O l 4e O2 g
Eo 0:401 vs: NHE
2H2 O l 2e ! 2OH- aq H2 g E 0:828 V vs: NHE

o
4:4
4:5
In the simplest form, these two oxidation and reduction reactions can occur over
metal electrodes (oxygen is produced at the anode, and hydrogen is produced at the
cathode), with current and voltage supplied by an external power supply. The
challenge of PEC water splitting is to create this external voltage and current
directly from converted solar energy in a monolithic device. The actual means to
do such a conversion can be accomplished in many different ways, which are
described in detail in the following sections.
In this chapter, several PEC device designs will be considered with respect to the
architecture of the components used, the management of different photo-absorbing
and catalyst materials, and the general operating principle that governs the synergy
of these materials. Furthermore, the focus of materials used will be inorganic,
meaning the light-absorbing compounds discussed will be semiconductors, which
have shown numerous applications in PEC devices that are able to achieve overall
solar water splitting, as opposed to molecular absorbers, which have shown poor
stability and conversion efficiency (to date), which has limited their applications in
practical devices.
4 Photoelectrochemical Cell Design, Efficiency, Definitions, Standards. . .
4.2
4.2.1
165
The Photoelectrochemical Cell

Cell Design
A photoelectrochemical reaction takes place in a photoelectrochemical cell. At the

very least, the PEC cell hosts the (photo)anode, the (photo)cathode, and electrolyte
solution. The major difference between a PEC cell and a (dark) electrochemical cell
is that one or more of the electrodes is photoactive, and thus needs to absorb light to
drive one or both of the chemical reactions associated with water splitting. Therefore, a PEC cell must have at least one transparent window in order to allow light to
penetrate the cell and be absorbed by one or both electrodes. Furthermore, the
anode and cathode must be in electrical contact, and so a conductive wire is needed
to connect the two electrodes, or the electrodes must be monolithically integrated
on opposing sides of an electrically conductive substrate, as shown in the design of
the artificial leaf (Nocera 2012). For the sake of clarity, the electrode (either
photoanode or photocathode) that absorbs light can be called the working electrode,
while the electrode that drives the opposing half reaction that is not light-activated
is called the counter electrode. In the case that both electrodes are photoelectrodes,
they can both be referred to as the working electrodes and are further clarified by
being referenced as the working photoanode and the working photocathode.
In an ideal case where bias-free water splitting can be achieved, the aforementioned materials are the only components of a PEC device that are needed to convert
solar energy and water into hydrogen and oxygen. However, for systems that
require an extra bias to drive the water splitting reaction, or for detailed diagnostics
of the mechanisms for achieving water splitting, more components are needed. For
example, to accurately examine the potential of the working electrode, a reference
electrode is needed. In addition, to maintain the electrolyte concentration, minimize
pH gradients, and aid in reactant/product mobility, magnetic stirrers and gas
circulation are necessary. An illustration of a PEC cell with the aforementioned
components is shown in Fig. 4.1a. If the two electrodes (working and counter) are
too close to each other, it is possible that the gas evolved at one surface may
contaminate or back-react with the reactants/intermediates of the other electrode.
Therefore, a membrane may be used to prevent crossover of gaseous products,
which directly allows hydrogen and oxygen to be evolved into their own separate
containers. Such a system would then require two separate circulation systems, and
an illustration of this configuration is shown in Fig. 4.1b.
These PEC cell designs are useful for an electrochemical system where the
working and counter electrode are spatially separated; however, many potential
PEC device designs favor the approach of a monolithic device that has the working
and counter electrode on one substrate. An illustration of a PEC cell design for a
monolithic system is shown in Fig. 4.2.
As can be seen from the previous figures, a working PEC cell used for overall
solar water splitting requires many components, each of which has a large variability in terms of materials that can be used, stability, and standardization.
166
W.A. Smith
Fig. 4.1 Illustration of the basic components of a photoelectrochemical cell where there is (a) a
single compartment for the working and counter electrode and (b) two compartments that separate
the working and counter electrodes by a membrane
167
Fig. 4.2 Illustration of the basic components of a photoelectrochemical cell with a monolithic
device combining the working and counter electrode onto one substrate
The following sections will briefly describe the different elements of an overall
PEC cell, and how they can be used to determine the performance and efficiency of
a solar water-splitting device.
Before mentioning the individual components, it is useful to illustrate an actual
working PEC cell in slightly more detail than in Figs. 4.1 and 4.2. A conceptual
design of a working PEC cell designed by van de Krol is shown in Fig. 4.3 (van de
Krol 2012).
This cell is made from PTFE and is custom-designed to fit a working electrode,
counter electrode, electrolyte, reference electrode, and transparent window to allow
solar irradiation to penetrate the cell and hit the photoactive electrode. In this
design, the sample (deposited on a flat and conductive substrate) is mounted on
the left side of the cell and makes an airtight seal with the body of the PEC cell.
There is a small chamber in the bottom of the cell to allow a magnetic stirrer to be
used, which can help distribute reactants and disperse products during electrochemical reactions. More details about this PEC cell, its design, and functionality can be
found in ref (van de Krol 2012).
168
W.A. Smith
Reference electrode
feedthrough
PTFE lid with feedthroughs for counter

and (quasi-)reference electrodes,
and gas circulation/bubbling
Holes for cell

alignment rods (4x)
R3
Cell body (PTFE)
Fused silica window

(50 mm)
Sample insert
Sample
R2
R1
Drain
Fig. 4.3 A practical PEC cell designed by van de Krol, used with permission from (van de Krol
2012)
4.2.2
Electrodes
A practical PEC device needs to be able to manage optical, electronic, and catalytic
functionalities all at the same time. For a true photoelectrochemical device, the
light-absorbing material should be immersed in the liquid, significantly increasing
the difficulty and complexity of cell design considerations. For example, depending
on the configuration of the device, the incident light may need to travel through the
water/electrolyte first, which can reduce the photon flux that is received at the
semiconductor surface, thus decreasing the possible maximum photocurrent that
can be obtained. This in turn decreases the potential solar to hydrogen conversion
efficiency (STH) of a practical device. Conversely, if the light does not need to go
through the electrolyte first, and instead goes through the back of the substrate,
different opto-electronic requirements for the substrate are required to maximize
the efficiency of the device. This section describes the different configurations that
the materials of a PEC device can have, and how the semiconductors and catalysts
can be arranged in different ways that can affect the stability, efficiency, and
practicality of a solar fuel system. The optimization required to achieve high STH
efficiencies with tandem device configurations, i.e. band gap matching, spectral
169
utilization, etc., will not be discussed in this chapter, as it is elaborated on in explicit

detail in Chap. 12. Furthermore, the separation of anode and cathode into separate
spatial components (i.e. a wired device) versus a monolithic-integrated device
(i.e. a wireless device) will not be discussed in this chapter, as it is also discussed
in Chap. 12.
4.2.2.1
General Considerations
The following discussion about PEC device design and considerations is focused on
semiconductor thin films and will not elaborate on particle-based systems that lack
an electrically wired configuration. Furthermore, the discussion will focus on
lab-scale devices and architectures used mainly for testing efficiencies of materials
and device configurations and not emphasize the up-scaling towards reactor and
industrial level designs, which have been excellently elaborated in several key
publications (James et al. 2009; Pinaud et al. 2013; Sathre et al. 2014). Finally,
the working principle for the different architectures will be briefly discussed in
order to speculate on the possible performance limitations of each configuration,
which in turn may be used to choose a different cell design or orientation in order to
obtain the maximum possible overall solar water-splitting efficiency of a device.
For both single-component photoanode and photocathode films (i.e. without
buried junctions which will be discussed later in this chapter), the light-absorbing
material needs to be deposited on a highly conductive substrate (the current
collector) that allows charges to be extracted or injected between the working
electrode and counter electrode. For an n-type photoanode, photogenerated holes
migrate towards the surface to perform water oxidation, and thus electrons should
flow through the bulk of the semiconductor, to the back contact, through a conductive wire, arrive at the counter electrode, and there reduce water/protons. For a
p-type photocathode, photogenerated electrons migrate towards the semiconductor
surface where they reduce protons/water, and holes must migrate through the bulk
of the material to the back contact, through a conductive wire, arrive at the counter
electrode, where they must oxidize water. In both cases, an ohmic contact is
required at the semiconductor/back-contact interface, and thus highly conductive
layers are typically used to form the top layer of the substrate before depositing a
photoelectrode. However, if the device requires light to be incident from the backside of the sample (i.e. light hits the substrate before the photoelectrode), the ohmic
contact must also be transparent to light. It is again important to note that these
considerations are only valid for a single-absorber system and do not hold if a
tandem absorber electrode is constructed as the light path would need to travel
through more than one light-absorbing material, and thus have different optimization criteria.
With the aforementioned requirements for an ohmic back-contact, which may or
may not be transparent, it is possible to find materials which fit such a specific
criteria. The most widely used materials for this application when it is necessary to
have back-side illumination (or for a tandem device) are transparent conducting
170
W.A. Smith
oxides (TCOs) such as F-doped SnO2 (FTO), In-doped SnO2 (ITO), and Al-doped
ZnO (AZO). These TCO layers have a relatively high conductivity (with respect to
typical semiconductor photoelectrodes) and allow a large amount of spectral
transmission so that the incident solar irradiation is maximized when it hits the
light-absorbing photoelectrodes. Typical conductivities (in S/cm) for FTO, ITO,
and AZO are 1 103, 1 104, and 7 103, respectively.
In addition to the requirements that are necessary for a transparent ohmic
contact, there are serious implications for the practical efficiency of a PEC device
by using front-side or back-side illumination. An example of these implications is
shown in Fig. 4.4, which illustrates the photogenerated charge carriers created for
an n-type photoanode. For the case of front-illumination (Fig. 4.4a), most of the
absorption in the photoanode will occur near the surface of the electrode, and as the
light is absorbed through the thickness of the material, less light reaches the back of
the electrode. The result of this is a greater density of photogenerated charge
carriers near the surface of the semiconductor than at the back of the electrode.
Near the surface, photogenerated holes are very close to the semiconductor-liquid
junction (SLJ), and thus the hole diffusion length does not need to be very long. On
the other hand, the photogenerated electrons created near the surface need to diffuse
through the bulk of the electrode to the back-contact where they are extracted and
transported to the counter electrode for hydrogen evolution. Therefore, for a
photoanode being subjected to front illumination, it is important that the electron
diffusion length is greater than or equal to the thickness of the films. Conversely, for
a photoanode illuminated from the back-side, as shown in Fig. 4.4b, there is a
higher density of photogenerated charge carriers closer to the back-contact than the
Fig. 4.4 Schematic illustration of a system where (a) light is incident on the semiconductor
surface first, i.e. front-side illumination, and (b) where the light is incident on the substrate side
first, i.e. back-side illumination
171
Table 4.1 Typical metal oxide photoelectrodes and their associated carrier mobility, carrier
lifetime, and diffusion length
Photoelectrode
material
Fe2O3
WO3
Cu2O
BiVO4
TaON
Ta3N5
Carrier mobility
(cm2 V2 s1)
0.5
10
6
0.044
0.01
0.07
Carrier lifetime (s)

3 1012
1~9 109
40 1012
40 109
1 103
1 103
Diffusion length (nm)

2~4
150~500
25
70
~31
~84
SLJ. Therefore, for this case, the electrons only need to diffuse a very short length to
reach the back-contact, and the holes must be able to diffuse through the bulk of the
material without recombining in order to reach the SLJ and oxidize water. These
two cases can be flipped for a configuration with a photocathode, where the
diffusion of charge carriers is opposite, i.e. the photogenerated electrons need to
reach the SLJ and the photogenerated holes need to diffuse to the back-contact.
The importance of this formalization is that the electron and hole diffusion length
and mobility is different for many materials. An example of the mobility of electrons
and holes for commonly used photoelectrode materials is shown below in Table 4.1.
These values are important because they can help to determine how thick a film
should be to balance the maximum absorption of light with the transport of
photogenerated charge carriers, and thus which type of illumination, i.e. frontside or back-side, should be used based on the diffusion length of the minority
and majority charge carriers.
In addition to the selection of back-contact material, the actual photoelectrode
materials must be fabricated and deposited on the substrate. The choice of materials
and device configuration is not straight forward, and a number of device geometries
and compositions have been reported in the literature. In general, it is possible to
de-couple or distribute the task of light harvesting and catalysis into different
materials that together make up a working photoelectrode. The litany of configurations of photoactive and catalytic materials can be divided into three main
categories: (1) photovoltaic cells with electrocatlayst layers deposited on top of
them, (2) photovoltaic cells with photoelectrode layers deposited on top of them,
and (3) a fully photoelectrochemical device with either/both a photoanode and/or
photocathode, i.e. one or both water oxidation and/or reduction are photo-driven
reactions. The working principles and outlook for each design configuration is
given in the following sections.
4.2.2.2
Photovoltaic Cell + Electrocatalyst (PV + EC)
The most straightforward approach to achieve efficient solar-driven water splitting

is to completely separate the light absorption and catalysis functions. This can be
172
W.A. Smith
done most effectively by having a PV panel to convert solar energy into electricity,
which can then be connected in series to an electrolyzer, which can perform the
water-splitting reaction. Such a configuration may be accomplished on a labscale
with the PV and electrolysis components very close to each other and with small
dimensions, (Luo et al. 2014; Cox et al. 2014). However, such an approach on a
large scale would require inverters to convert the DC current generated by the solar
cells to usable current for the electrochemical cell. This system integration may be
the simplest from a practical standpoint and can use already developed off-the-shelf
PV components with industrial scale electrochemical cells. However, as of now this
approach is not cost-effective, as the cost of the total system integration would
produce hydrogen that is not competitive with the current price of fossil fuels
(James et al. 2009; Pinaud et al. 2013; Sathre et al. 2014).
This approach actually makes the production of oxygen and hydrogen from
water completely separate from any light absorption and dependent only on a
current/voltage supply. This allows the separate optimization of solar to electricity
conversion, as well as the dark catalytic reactions for the OER and HER, for which
recent benchmarking of OER and HER catalysts has been assembled for acidic and
alkaline environments (McCrory et al. 2013, 2015). The advantage of such a system
is that any source of electricity (preferably renewable) can be used to power the
electrolysis reactions. One logical extension of this device configuration is to use
large-scale utility PV panels connected to grid-powered electrolyzers. Using such a
large scale system, it is possible to tackle the terawatt scale demand the world has
and is thus the most likely and technologically advanced way to store solar energy
through water splitting. However, industrial electrolyzers are run at large current
densities (hundreds of mA/cm2) and require precious metal catalysts that are only
stable under these conditions and easily corrode when their current/voltage source
is removed. This is a strong set-back if this system is to be used with solar energy as
the input, as sunlight is intermittent, and thus the electrolyzers would experience
significant downtime and thus corrode. Furthermore, since the light-absorbing PV
units are not immersed in the electrolyte, this is technically not a photoelectrochemical device and is only an electrochemical cell powered by renewable electricity. For these reasons, this device architecture will not be discussed in more
detail in this chapter.
While a large-scale PV cell coupled to an electrolyzer has several potential
disadvantages as listed above, the direct integration of the two components into a
monolithic device offers a potential solution to the combination of materials. An
overall schematic for such a device using a 3-jn a-Si PV cell is shown in Fig. 4.5a,
with the associated theoretical electronic band diagram given in Fig. 4.5b.
In this configuration, a multi-junction PV cell is coated with an ohmic contact on
each side in order to prevent corrosion and offer excellent electronic charge
transfer. A multi-junction PV cell is shown here, as it is more likely to provide
the necessary photovoltage to drive the water-splitting reaction, taking into account
the thermodynamic potential for water splitting plus necessary overpotentials, when
compared to a single junction solar cell. On top of the ohmic contacts, a dedicated
oxygen evolution catalyst (OEC) and hydrogen evolution catalyst (HEC) are
173
Fig. 4.5 A cartoon representation of a wireless monolithic PV-electrolysis cell (a), and the
associated electronic band diagram for a potential triple junction PV device that powers the
water-splitting reaction, where OEC is the oxygen evolution catalyst and HEC is the hydrogen
evolution catalyst
deposited, which carry out the respective redox reactions efficiently. Such a device
architecture has been successfully used extensively in the literature (Appleby
et al. 1985; Sakai et al. 1988; Lin et al. 1989; Rocheleau et al. 1998; Khaselev
et al. 2001; Licht et al. 2001; Reece et al. 2011). One distinct advantage of such a
system design is that it is possible to combine state-of-the-art materials for each
component, i.e. you can potentially use a high performance PV cell with excellent
optoelectronic properties (high Voc, high Jsat), and couple this to high performance
hydrogen and oxygen evolution catalysts (low overpotential, high turnover frequency). However, one practical limitation of such a device is that due to the
architecture, the light-absorbing PV cell is buried beneath both hydrogen and
oxygen evolution catalysts, and thus the incident solar irradiation must first pass
through one of the catalyst materials. This may introduce more optical losses in the
overall system, since light can either be reflected or absorbed by these extra layers.
This is particularly troubling since many of the state-of-the-art HECs are made
from metallic precious metals such as Pt and Ir, and the state of the art OECs
become dark when a potential (in this case a photovotlage) is applied (Bendert and
Corrigan 1989; Corrigan and Knight 1989; Conell et al. 1992; Trotochaud
et al. 2013). Therefore, the combination of materials to make a monolithic
PV-electrocatalyst device is not as straight forward as simply connecting efficient
PV cells and electrocatalysts, but their optical properties must be taken into
account. A technical summary of the effect of the different configurations of the
light path going through either the OEC side or the HEC side on the photoelectrochemical performance was recently developed by Seger et al. (Seger et al. 2014).
The true implications for the described PEC device configurations are more closely
tied to the optimization of a tandem PEC device, with multiple absorbing materials,
and are discussed in further detail in Chap. 12.
A slight modification of this approach can be achieved by changing the arrangement of the PV cell, while still maintaining the same architecture, i.e. having a
horizontal protected multi-junction PV cell, and having a separate dedicated HECs
174
W.A. Smith
Fig. 4.6 (a) A schematic of a side-by-side PV-EC system, and (b) theoretical maximum STH
efficiencies as a function of the PV band gap energy, adapted reference (Jacobsson et al. 2015)
and OECs spatially separated. In this device configuration, PV cells can be placed
side-by-side and connected in series, as shown in Fig. 4.6.
This side-by-side PV system has two practical advantages compared to the
monolithic device previously described. Since this architecture places the PV
cells side by side, the optical absorption by each PV junction does not interfere
with the others ability to absorb light, i.e. there is no parasitic light absorption for
the light going through consecutive layers as is necessary in the device shown in
Fig. 4.5. This opens the potential for using series-combined PV cells that are each
optimized for the entire solar spectrum, removing the requirement to have buried
PV junctions that are optimized for the transmitted spectra of light that pass through
top layers, i.e. having to match top-cell and bottom-cell band gap energies, as
described in Chap. 12. The drawback of such a device architecture is that it extends
the solar irradiation surface area, thus making the current density (i.e. photocurrent
density) that travels to the catalyst surface area much more dilute. The limitations of
such a device configuration were discussed by Jacobsson et al. (2015), who found
the theoretical potential of such devices can be up to ~20 % STH, slightly lower
than a traditional monolithic device (see Chap. 12). An important conclusion of this
study showed that a high theoretical STH conversion efficiency is obtainable in this
architecture, implying that this device configuration is still valuable to be explored.
For the consideration of the forthcoming devices, it is important to note that a
PV-electrolysis configuration has no semiconductor liquid junctions (SLJ), because
the light-absorbing semiconductor is not in direct contact with the electrolyte. This
is an important feature to note, since the interfacial band edge energetics at the
semiconductor-liquid junction differ significantly than those determined by ohmic
contacts and Schottky barriers that are associated with PV-electrolysis devices. It is
also important to note here that the band edge positions of a PV cell are irrespective
of the applied potential/redox potentials in the solution (determined by
electrocatalysts in contact with the electrolyte).
In order to facilitate understanding of how this and other device configurations
work, it is useful to look at the operating mechanisms of this system. In particular,
by combining the current vs. potential plots for the two separated systems (PV and
electrocatalyst), it is possible to extract an operating point for the combined system,
175
Fig. 4.7 A hypothetical

current versus potential plot
of two solar cells and one
oxygen evolution catalyst,
with a reference water
oxidation potential shown
to illustrate that an oxygen
evolution catalyst must
have more potential applied
than the thermodynamic
potential of this reaction
and from this estimate the potential and limitations of a best case scenario,
forming a future prospective for further technological and scientific development.
The JV curves of two hypothetical PV cells are shown in Fig. 4.7 (black dashed
lines), where the short circuit density (Jsc) and open circuit potential (Voc) are shown.
Typically, JV curves of PV cells are measured in a two-electrode configuration,
since they are solid state devices that are connected by Ohmic contacts, and thus the
potential is directly measured across the donor and acceptor regions of the photovoltaic device. In the same figure, a JV curve of a hypothetical oxygen evolution
catalyst is shown (gray dashed curve) with its corresponding onset potential (Von) for
catalysis. Where the two curves intersect is the operational point (Vop) of the
PV-powered electrolysis device. For reference, the water oxidation potential is
shown (while its real value is 1.23 V vs. RHE, here the figure is illustrative and
thus does not represent the actual potential value, only to demonstrate it is to the left,
i.e. at a lower potential than the JV curve of the electrocatalyst.). It is again useful to
note that for an electrochemically derived JV curve (either for an electrocatalyst or
a photoelectrocatalyst), the potential is typically measured in a three-electrode
configuration, where the potential of the working electrode is measured with respect
to a third (reference) electrode with a known redox potential. More details about
two-electrode and three-electrode measurements will be discussed in Sect. 4.4. From
this figure, several important limitations can be extracted. First, from the
electrocatalyst side, a dark (i.e. not photocatalytically active) electrocatalyst can
never have an onset potential less than 1.23 V vs. RHE (VH2O/O2), because this is the
thermodynamic equilibrium potential of water oxidation. This illustrates that the
dashed gray line corresponding to the JV performance of the OEC will never
become more cathodic (i.e. be to the left of) to the water oxidation potential of
1.23 V. The implications of this are that the PV cell used to provide current and
potential to the OEC must then satisfy the requirement of having a Voc higher than
(i.e. to the right of) the water oxidation potential. In practice, the actual state-of-theart OEC materials still require a minimum of 200~300 mV overpotential to drive this
reaction (McCrory et al. 2013, 2015), meaning the Voc of a practical PV must be at
176
W.A. Smith
least 1.4~1.5 V. in order to operate in a bias-free device. This puts a strict limit on the
performance characteristics of the PV cell, which can influence the potential applicability of an efficient, scalable, and cost-effective device. First, there is an inherent
tradeoff for PV cells between the Jsc and Voc, i.e. the higher the Jsc the lower the Voc
and vice versa. This means for a PV cell to have a Voc higher than 1.23 V, the Jsc will
be reduced and thus have a lower performance. In addition to the current density
limitations that a high Voc places on a PV cell, the ability to achieve such a
performance in a stable and cheap material may be limited. For example, the leading
PV materials that can achieve such high current densities and large Vocs are either
made from very expensive materials (GaAs) or are unstable (perovskites), and thus
may not be practical for a cost-effective system to produce solar hydrogen. At the
moment, perovskite PV cells have gained significant attention due to their rapid
growth in cell efficiency (over 20 % as of 2015); however, the materials are
inherently unstable and thus the use of such a material class, at the moment, seems
unlikely. The most practical material that can be used to balance cost and efficiency
is silicon, with crystalline silicon, c-Si, having a band gap energy of 1.1 eV and
amorphous silicon, a-Si, having a band gap energy of 1.8 eV. However, the overall
efficiency of single and multi-junction Si solar cells may be limited due to the low
Voc obtainable for these materials, thus limiting the potential Vop. While this
seemingly puts many restrictions on a PV + electrolysis cell, there are other device
configurations that can have a more beneficial JV performance operating point,
which will be discussed in the next sections.
4.2.2.3
Photovoltaic Cell + Photoelectrode (PV + PEC)
While PV-electrocatalyst devices offer the ability to directly combine state-of-theart PV materials with state-of-the-art electrocatalyst materials in a straight forward
manner, there are many possible limitations of using such a device in a practical
application. A variation of this architecture is to couple a PV cell with a
photocatalyst material that can directly photo-drive either the water oxidation or
reduction reaction. A sketch of a wireless monolithic PV/PEC system with a
photoanode (i.e. a photocatalyst driving the water oxidation reaction) is shown in
Fig. 4.8a, with the associated band diagrams shown in Fig. 4.8b.
This PV/PEC system has several advantages and disadvantages when compared
to a direct PV-electrocatalyst system. From a fabrication and cost perspective, this
architecture is simpler, and thus possibly can be more promising for upscaling to a
large area device. This is because the single photoanode layer (in this particular
configuration) replaces one p-i-n PV junction and tunnel layer. From a manufacturing point of view, this means depositing one layer instead of four, which obviously
can reduce overall device costs, provided the photoanode material and fabrication
process are cheaper than the 4-layer p-i-n and tunnel junction layer depositions.
Furthermore, since the photoanode (in this case, though the same is true for the
opposite case with a photocathode) is in contact with the electrolyte, a semiconductor liquid junction is formed. This means that the interfacial band edge
177
Fig. 4.8 A cartoon representation of a wireless monolithic PV-PEC cell with a double-junction
PV cell attached to a photoanode to drive the water oxidation reaction (a), and the associated
electronic band diagram for a potential double-junction PV/PEC device that powers the watersplitting reaction
alignment of the Fermi-level in the photocatalyst should, in principle, align with the
relevant redox reaction potentials. In an ideal case, this alignment can happen
directly with no losses or added overpotentials, but in practice, overpotentials
exist due to kinetic-driving forces required to carry out the 4-step water/OH
oxidation and 2-step water/proton reduction reactions, as well as the presence of
electronic surface states that may pin the electronic band energies at potentials less
than the highest achievable photovoltage.
While the band edge positions of a PV-cell are not dependent on the redox
potentials in the solution, for a photoelectrode that absorbs light and drives a
chemical reaction, the valence band (Ev) and conduction band (Ec) positions must
be favorable relative to the water oxidation and reduction potentials. In particular,
the valence band should be lower (more positive) than the oxygen evolution
potential, and the conduction band should be higher (more negative) than the
hydrogen evolution potential. Ideally, a single material could be used to drive the
overall reaction, with conduction and valence bands that straddle the hydrogen and
oxygen evolution potentials; however, such a material has not yet been found or
developed to an efficient device. More on the practical utilization of single absorber
materials is described in Chap. 12.
Similar to the previous section, it is useful to compare the JV characteristics of
the different components of this system to see its operational principle, and its
inherent advantages and disadvantages. A current vs. potential plot is shown below
in Fig. 4.9, where the photocurrent is shown for the photoelectrode (in this case a
photoanode) in the dashed gray, and the JV characteristics of the buried junction
PV are shown in the dashed black line.
Similar to the PV-electrolysis case, the intersection of the two JV curves is the
operational point of the device. The most striking difference to the previous case is
that here the intersection point can come at a lower potential than the water redox
potential, which gives more flexibility in terms of system optimization. In particular, using this PEC/PV approach offers significant flexibility in lowering the Vop of
a practical device, and thus may offer a more realistic pathway towards a high
178
W.A. Smith
Fig. 4.9 A hypothetical

JV plot of a photoanode
and a series connected
PV cell
efficiency, low-cost PEC device, as many benchmark device performances have

been shown using this device configuration (Morisaki et al. 1976; Khaselev and
Turner 1998; Miller et al. 2005; Park and Bard 2006; Arakawa et al. 2007; Gaillard
et al. 2010; Brillet et al. 2012; Abdi et al. 2013; Han et al. 2014).
4.2.2.4
Dual Photoelectrodes (Photoanode + Photocathode, i.e. PEC)
While a PV-PEC device offers some advantages and disadvantages compared to a

PV-electrocatalyst system, further consideration can be applied to a fully PEC
system composed of a photoanode and/or photocathode. In such a device configuration, the water/OH oxidation and water/H+ reduction reactions can both be
photodriven, giving the system 1 or 2 semiconductor liquid junctions. In a 2 SLJ
system, compared to a PV-PEC system, this may provide more losses at the SLJ
interfaces due to catalytic overpotentials for each reaction. In addition, this configuration (2 photoelectrodes) requires light to pass through the electrolyte in all cases,
giving rise to further optical losses that may limit the potential efficiency
achievable.
From an operational point of view, the current matching for a fully PEC
photoanode/photocathode system is identical to the PV-PEC system, as shown in
Fig. 4.10. The main difference is that the fully PEC system has 2 SLJs, and thus the
obtainable operating current is limited by the photovoltages that can be obtained by
both the photoanode and photocathode, which in practice will be smaller than those
of a PEC and PV material. In short, the PV material will obtain its maximum Voc by
having ohmic contacts on either side, while the PEC is limited by a SLJ, which
requires equilibration of the Fermi level and dominant redox potential associated
with either desired chemical reaction. From a practical perspective, this may limit
the overall potential of such a system, which is shown by the poor demonstrated
efficiencies in the literature (Nozik 1976; Kainthla et al. 1987; Mor et al. 2008; Sato
et al. 2011; Arai et al. 2013; Bornoz et al. 2014).
179
Fig. 4.10 A cartoon representation of a wireless monolithic PEC cell with a photoanode to drive
the water oxidation reaction and a photocathode to drive the water reduction reaction (a), and the
associated electronic band diagram for a potential single-junction PEC/PEC device that powers the
water-splitting reaction
4.2.3
The Electrolyte
The composition of the electrolyte used in a PEC cell is essential to the performance
and stability of the overall device. For the considerations of this chapter where we
only discuss PEC water splitting, the electrolyte solution will be composed of liquid
water with different solvated ions. Water itself is a poor conductor, so it is necessary
to dissolve charged ions to aid in the charge transfer process between the working
and counter electrode. Many considerations need to be accounted for in using a
particular ionic species in an electrolyte including the materials stability, the ionic
conductivity, and the diffusion of each ion through a potential membrane. The role
of the electrolyte, which is an ionic liquid solution, is to transfer charge between the
surfaces of the working and counter electrodes. Positive charge is passed through
protons (H+), while negative charge is passed through hydroxide ions (OH). In an
aqueous solution with the standard used ionic species, H+ and OH have the highest
limiting ionic conductivities of 349.8 and 197 (104 1 mol1 m2), respectively.
Other commonly used cations such as K+ and Na+ have lower limiting ionic
conductivities of 73.5 and 50.1 (104 1 mol1 m2), respectively, while commonly used anions such as Cl and SO42 also have lower limiting ionic conductivities of 76.4 and 162 (104 1 mol1 m2), respectively. A table of commonly
used acid, base, and neutral solutions of various concentrations are shown in
Table 4.2, where the conductivity of the electrolyte, , the electrolyte resistance,
RE, and potential loss at 5 mA/cm2 are given for each electrolyte composition.
While the measured ionic conductivities for the different ionic species may seem
high, relative to the conductivity of electrons through a conductive wire (for copper,
conductivity, , ~ 6 107 S/m), they are several orders of magnitude smaller. The
relatively low ionic conductivities can lead to large ohmic losses, which increase
necessary overpotentials to drive the water splitting half reactions, thus decreasing
the overall device efficiency. The comparison of the conductivity of a metal wire
and an ionic solution is important when considering the design of a monolithic PEC
device. For example, if the working and counter electrode are spatially separated, as
180
W.A. Smith
Table 4.2 Typical electrolyte compositions and acidity/pH with the associated conductivity,
resistances, and potential losses at 5 mA/cm2, adapted from (van de Krol 2012)
pH
Neutral
Neutral
Acid
Neutral
Base
Electrolyte
composition
Distilled water
Purified water
0.5 M K2SO4
1.0 M H2SO4
3.5 M H2SO4
0.1 M NaCl
0.5 M NaCl
1.0 M NaCl
0.1 M KOH
0.5 M KOH
1.0 M KOH
(1m1)
103~104
~5.5 106
6.2
36.6
73.9
1.07
3.8
7.44
2.26
10.7
20.1
RE ()
105~106
~18 106
16
2.7
1.4
93
26
13
44
9.3
5.0
Vloss @
5 mA/cm2 (mV)
1
1
81
14
7
467
132
67
221
47
25
T C
20
25
20
18
18
18
18
18
18
18
20
shown in Fig. 4.1a, b, the distance for ionic diffusion, and thus the ionic conductivity is decreased compared to the integrated monolithic device shown in Fig. 4.5.
The detailed comparison between a wired versus wireless PEC device is discussed
explicitly in Chap. 12.
In addition to the bulk composition of an electrolyte designed to transfer positive
and negative charges, the addition of a buffer to an electrolyte can have a very
positive effect. For example, when OH or H+ are consumed at an electrode/
electrolyte interface, the local concentration (i.e. pH) of charged ions in the solution
becomes slightly more acidic or basic, respectively. This small change can alter the
kinetics and possibly thermodynamics of the desired chemical reaction. Therefore,
buffer salts are used to react with the increased/decreased pH layers in order to
maintain a steady pH balance in the entire solution, but most importantly near the
electrode/electrolyte interface. Common buffers are phosphate (KH2PO4/K2HPO4)
and borate (H2BO3/HBO3), which maintain a solution pH at ~7 and ~9,
respectively.
The composition of the electrolyte is important in determining the ionic transport in the PEC cell, but is also critical in determining the stability of the working
and counter electrodes used during operation. In particular, the materials used in a
PEC cell should not corrode during prolonged exposure to light and the electrolyte.
A detailed chart of typical PEC photoelectrode materials and their associated selfreduction (black horizontal lines) and self-oxidation (dark grey horizontal lines)
potentials have been accumulated by Chen and Wang (Chen and Wang 2012),
shown in Fig. 4.11.
In view of this reference, it is clear that the electrolyte for a PEC cell must not
only be chosen for its favourable charge/ionic transport properties between the
working and counter electrodes, but must also be favourable for the long-term
stability of the electrode materials. Therefore, the (photo)electrodes and electrolyte
181
Fig. 4.11 The electronic band diagram and associated self-reduction and oxidation potentials for
selected semiconductor materials at pH 0, figure taken with permission from reference (Chen
and Wang 2012)
solutions must be chosen and carefully selected in tandem and never considered
irrespective of each other.
4.2.4
Ion Exchange Membranes
To avoid mixing of the produced fuel (hydrogen or carbon-based products) with

oxygen gas, ion exchange membranes are often applied to separate the anodic and
cathodic electrolyte in proposed solar fuel designs. Three types of ion exchange
membranes are discussed: proton exchange membranes, anion exchange membranes and bipolar membranes. In addition, also membrane-less designs that
avoid mixing produced hydrogen and oxygen gasses have been suggested in
literature and will be discussed briefly.
4.2.4.1
Proton Exchange Membranes
Based on the traditional fuel cells and electrolysers, a proton exchange membrane
(PEM) can be used in solar fuel devices to allow the transport of H+ from the anode
to the cathode (Haussener et al. 2012; Roy et al. 2010). The most commonly used
proton exchange membrane is Nafion, which is known for its high conductivity,
high chemical stability, and optical transparency. Other proton exchange
182
W.A. Smith
membranes have been proposed and investigated (Hickner et al. 2004; Peckham
and Holdcroft 2010), but are still less widely used.
Two issues related to proton exchange membranes can be identified for the use in
solar fuel devices. First, the price of these membranes is high, which is in particular
an issue because the low current density requires a large area. Second, in contrast to
what the name suggests, proton exchange membranes also allow the transport of
other cations than protons (Chae et al. 2007), albeit with lower conductivities.
Consequently, when other cations (e.g., K+ or Na+) dominate the concentration of
protons, which is usually the case at pH > 1, these cations partly account for the
charge transport through the membrane, while protons are consumed at the cathode
and produced at the anode. In the long term, the pH at the cathode will increase
while the pH at the anode will decrease, which increases the required voltage for
water splitting (polarisation) (McKone et al. 2014). Modestino et al. (Hashemi et al.
2015) have shown that partly mixing the anodic and cathodic electrolyte can limit
this effect to a single pH unit, with a minor compromise in gas purity.
4.2.4.2
Anion Exchange Membranes
As an alternative for the proton exchange membranes, anion exchange membranes

are proposed for solar fuel devices as well, which only allow the transport of anions
such as OH (McKone et al. 2014). In order to facilitate OH transport rather than
transport of other anion species, the use of anion exchange membranes is in
particular attractive for alkaline electrolytes, which matches the activity of earth
abundant oxygen evolution catalysts (McCrory et al. 2015). For near-neutral solutions, anion species other than OH are transported as well, which creates additional voltage loss after multiple hours of operation (Hernandez-Pagan et al. 2012),
similar to cation or proton exchange membranes. In addition to that, anion
exchange membranes suffer from limited chemical stability in strongly alkaline
environments, lower conductivity (although that may not be an issue at current
densities 10 mA/cm2), and limited selectivity leading to cation crossover instead
of OH transport (Varcoe et al. 2014; Hickner et al. 2013). Hence, the development
of stable, selective, and possibly transparent anion exchange membranes is an
on-going challenge for the solar fuel development.
4.2.4.3
Bipolar Membranes
Electrolyte restrictions for the stability and activity of photoelectrodes and (co-)
catalysts limit the options of an integrated practical solar fuel device. To enlarge the
compatibility of (photo-)anodes and cathodes, a bipolar membrane (BPM) can be
used to separate the anodic and cathodic electrolyte. A bipolar membrane dissociates water into H+ and OH due to the two-layered ion membrane structure, which
allows maintaining a different pH at either side of the membrane (Simons 1993).
Compared to the other ion exchange membranes, the use of such membrane
183
provides the additional advantage of using different (stable) pH at either electrode.

This idea has been explored for dark electrolysis (McDonald et al. 2014; VargasBarbosa et al. 2014), in which water splitting with a cathode at pH 0 and an anode at
pH 14 did show no increased voltage with respect to the traditional case at a single
pH without a bipolar membrane. Recently, a BPM has been used in a photodriven
system consisting of a BiVO4 photoanode in pH 7 or 14 and a Pt cathode in pH 0
(Vermaas et al. 2015). The milder conditions, i.e. pH 7 versus pH 0, yield insignificant potential losses after 80 h of operation at current densities estimated to be
close to those needed to be produced in large-scale solar fuel devices.
4.2.4.4
Membrane-Less Systems
To avoid costs for membranes and to avoid polarisation over the membrane at nearneutral pH, membrane-less solar fuel systems have been proposed as well. Examples of membrane-less system with proven separation of hydrogen and oxygen
gasses are based on mesh electrodes with divergent convective flow (Gillespie
et al. 2015) or devices with fast tangential water flow along plate electrodes
(Hashemi et al. 2015). Although the latter system offers promising low hydrogen
and oxygen crossover (only a few percent), only microscale systems have been
tested as of yet. Similar for all membrane and membrane-less designs, the type of
system strongly depends on the electrode and catalyst requirements. Hence, as no
consensus is achieved for an integrated design for solar fuels, the options for one of
the mentioned membranes or membrane-less designs are all open for development.
4.3
Measurement Protocols
While the previous section describes the components and configurations of PEC
cells, it is also important to have well-defined protocols for measuring the performance and efficiency of PEC materials and systems. This is most important to aid in
the comparison of materials and devices made in different laboratories in different
countries around the world. Therefore, several performance benchmark metrics are
described in the following section, along with standard measurement protocols and
equipment so that the performance and efficiency of PEC materials and devices can
be normalized across the field.
4.3.1
Simulated Solar Irradiation Measurements
The most obvious measurement to consider for standard protocols is how to observe
the performance of a photoelectrode under solar irradiation. While the solar spectrum is constant from its source 93,000,000 miles away, there is a variance in the
184
W.A. Smith
Fig. 4.12 The AM 1.5 global solar spectrum with the indicated areas that correspond to the light
energy of 1.23 eV (dark grey) and 2.0 eV (light grey), which indicate the water splitting potential
and theoretical potential needed to drive actual water-splitting including losses, respectively
location where you measure its power, which also depends on the time of day and
season you are measuring. Therefore, a normalized standard solar spectrum and
power density has to be introduced in order to have a metric by which to standardize
materials performance. Such a standard has been used extensively for decades in
the photovoltaic field, and the same conditions are applied to the PEC field. The
agreed upon standard metric for simulated solar irradiation is global air mass 1.5
(AM 1.5), as shown in Fig. 4.12.
This illumination source must be calibrated in each lab by means of a photodiode
to ensure that the spectral distribution and power density is closely related to the
specifications. An extensive comparison between light sources and their specifications has been organized by R. van de Krol (2012), which the readers are guided for
reference.
For practical purposes, solar irradiation measurements are generally used while
performing linear sweep or cyclic voltammetry measurements, where the photocurrent density is measured as a function of applied potential. The information
gained from such a measurement is enormous as it can dictate the flatband potential,
saturated photocurrent density, and fill factor of a photoelectrode. An example of a
typical linear sweep voltammogram for a photoanode (BiVO4) and a photocathode
(a-SiC) is shown in Fig 4.13a, b, respectively. These materials and figures are used
to show the general trends for each class of material, i.e. to show that photoanodes
produce a positive (photo)current density when a positive potential is applied, and
that photocathodes produce a negative (photo)current density when a negative
potential is applied. For the following sections, the BiVO4 photoanodes were
deposited by a spray pyrolysis technique (as detailed in Abdi et al. 2013), and the
185
Fig. 4.13 Typical photocurrent vs. voltage plot for (a) an n-type BiVO4 photoanode and (b) a
p-type a-SiC photocathode
a-SiC photocathodes were grown by plasma-enhanced chemical vapour deposition

(PECVD) (as detailed in Digdaya et al. 2015).
In these plots, several interesting features can be observed. For each figure, the
dotted line represents the dark current, which is the current measured at different
potentials when no light is incident on the photoelectrodes. If a photoelectrode
shows any dark current at a given potential, it is usually a sign of corrosion, and thus
instability. Therefore, each of the dotted line plots in Fig. 4.13 indicate that the
materials are stable (i.e. do not corrode) within the potential range they are swept. It
is also interesting to observe that when each of the samples is illuminated by backand front-side illumination, the photocurrent generated shows different trends. For
example, with the BiVO4 photoanode, there is a higher photocurrent generated
when the sample is illuminated from the front-side, while for the a-SiC photocathode, there is a higher photocurrent generated when the sample is illuminated from
the backside. These measurements can be an indication of the performance limiting
photo-generated carrier diffusion length (see Sect. 4.2.2.1).
To clearly understand how much of the current density is due to the absorption
and conversion of sunlight (i.e. photocurrent), it is necessary to make measurements
under solar irradiation and in the dark. If the current density in the dark is subtracted
from the current density under illumination, this is called the photocurrent density,
where the density term applies to the areal coverage of the photoelectrode (usually
in cm2 for laboratory measurements). One approach to make such a plot is make
two (or more) separate measurements and plot them on the same axis as shown in
the dashed line of Fig. 4.13. An alternative is to make a chopped illumination
measurement, where a timed light-chopper is placed in between the light source and
the photoelectrode in timed intervals to give a single measurement profile that
alternates between light and dark measurements during a potentiodynamic sweep,
as shown in Fig. 4.14a for a BiVO4 photoanode and (b) for an a-SiC photocathode.
This measurement clearly shows that when light is able to reach the
photoelectrode, there is a sharp increase in the current density, and when the light
186
W.A. Smith
Fig. 4.14 Typical chopped illumination photocurrent vs. voltage plot for (a) an n-type BiVO4
photoanode and (b) a p-type a-SiC photocathode
path is blocked, there is a sharp decrease in the measured current density, which
relates directly to the dark current measurements as shown previously in Fig. 4.13.
Several important pieces of information can be extracted from both the JV
curves (shown in Fig. 4.13) and the chopped illumination curves (shown in
Fig. 4.14). When sweeping anodically/cathodically for photoanodes/photocathodes, the potential where the photocurrent generation begins is called the onset
potential, Von. According to Fig. 4.13, the Von for BiVO4 is ~0.6 V vs. RHE, while
the Von for a-SiC is ~ 0.8 V vs. RHE. While the Von are similar for the two materials,
it is important to again note that the trends are different for photoanodes and
photocathodes. In particular, the Von for BiVO4 implies that photocurrent will
begin to increase at potentials more positive than Von, while for the a-SiC photocathode the photocurrent generation will increase at potentials more negative than
Von. In addition, at potentials much larger than Von (more positive for photoanodes,
and more negative for photocathodes), the photocurrent density eventually saturates
at a maximum value, called the saturated photocurrent density, Jsc. Similar to the
PV-field, the slope of the JV curve as it moves from Von to the Jsc can give an
indication of the electronic properties and strength of the semiconductor used.
However, unlike in the PV-field where this fill factor is determined solely by the
intrinsic bulk properties of the semiconductor and not limited by the ohmic contacts
where charge carriers are extracted, for PEC materials, the fill-factor is determined by the SLJ, where electrons/holes are less easily exchanged due to poor
kinetics and the associated overpotentials. This is observed in the relative large
amount of potential that is required to reach Jsc after the Von (for the aforementioned
BiVO4 this potential is > 1.5 V vs. RHE, while for the a-SiC photocathode this
potential is > 1.2 V vs. RHE).
4.3.2
187
Determining the Flatband Potential
Similar to the Von mentioned in the previous section, an essential component of a

PEC material is the so-called flatband potential. The flatband potential, as indicated
by the name, is the potential at which no band bending occurs at the SLJ, and thus
the conduction and valence bands are flat, as shown in Fig. 4.15a for an n-type
photoanode. The actual potential is measured in a three-electrode configuration and
is defined as the potential between the Fermi level of the semiconductor and the
reference electrode.
These flatband conditions do not hold when a potential is applied that is greater
than the Vfb, and/or when the photoelectrode is illuminated, as shown in Fig. 4.15b.
In this case, the Fermi level of the photoanode is brought down below the previously determined flatband potential, either by the addition of an external bias, or by
the relative change in the electron-hole concentration due to illumination which
drives photoelectrocatalysis at the SLJ.
In order to actually measure the flatband potential, the most powerful technique
is impedance spectroscopy (IS), or more specifically, Mott-Schottky analysis
(Klahr et al. 2012). Using this technique, the capacitance of the space charge
layer, CSC, is measured, and 1/CSC2 is plotted against the applied potential, as
shown in Fig. 4.16 for a thin film of TiO2 grown by ALD (Digdaya et al. 2015).
From this plot, a linear slope can be made through the measured inverse capacitance
squared, and where the linear regression crosses the x-axis is the flatband potential.
In particular, the plot and subsequent linear region can be extrapolated from the
Mott-Schottky equation, given as:
Fig. 4.15 Semiconductor electronic band positions for (a) the flatband condition, and (b) with an
applied potential greater than the flatband potential, and illuminated
188
W.A. Smith
Fig. 4.16 A representative

Mott-Schottky plot for a
TiO2 film grown by atomic
layer deposition (ALD) on
an FTO substrate at 150 C,
used with permission from
(Digdaya et al. 2015)

1
2
kT
V

V

app
fb
e
C2SC 0 r eN D A2
4:6
The change of the flatband as a function of the pH of the electrolyte has been
found to be especially pronounced in metal oxide photoelectrodes, though it may
also hold for nonoxide semiconductors as well.
4.3.3
Evolved Gas Quantification
To ensure faradaic efficiency in photoelectrodes and to measure this in absolute

terms for dispersed photocatalyst materials, it is absolutely necessary to measure
and quantify the evolved products for the water-splitting reaction, i.e. to quantify
the amount of oxygen and hydrogen evolved. Similar to previous sections, here we
will not discuss the quantification of other solar fuel products, such as those made
from CO2 reduction, which may vary significantly, as it has been recently shown
that up to 16 gaseous and liquid products may be formed from a single CO2
electroreduction reaction (Kuhl et al. 2012). Therefore, the discussion in this
section only deals with the quantification of hydrogen and oxygen gases from
solar water splitting devices.
In order to determine the faradaic efficiency of either the hydrogen evolution or
oxygen evolution reactions, it is necessary to know three primary characteristics of
a film; (1) the active surface area of the catalyst, (2) the amount of current density
passing through the electrode, and (3) the number of moles of hydrogen/oxygen
produced as a function of time. If you can obtain all three of these criteria, then it is
possible to accurately describe how much of the current density measured goes to
189
the desired reduction/oxidation reaction, and how much goes to other processes
(i.e. side-reactions, back-reactions, corrosion, etc.). However, obtaining an accurate
estimation of the active sites in a chemical reaction, or even the amount of active
surface area, especially for a nanostructured (photo)electrode (Osterloh 2013), may
be very difficult to obtain. Therefore, in general, the actual surface area used in most
reports for semiconductor photoelectrodes is the projected surface area, or the
amount of area of the electrode exposed to the electrolyte, and does not include
nano-, micro-, or other sized features in the determination of the active surface area.
Thus, it may even be harder to compare current densities of different semiconductor
photoelectrodes, especially comparing planar electrodes to nanostructured
electrodes.
Furthermore, a large difference may be seen from making either static or
dynamic measurements of current density/gas production, and thus it is suggested
to make static voltage/current density measurements for more accurate measurements to allow for a more controlled production of oxygen/hydrogen. Using a fixed
potential and measuring the (photo)current density over time can also be a good
way to show stability/instability, as the current density will decrease if the sample is
unstable and generally remains constant if the system is stable (though the current
could also remain constant if there is a constant corrosion process).
4.4
Efficiency Definitions
In order to quantify the performance and efficiency of PEC materials and devices, it
is necessary to have well-defined benchmark metrics of assessment. Many reports
list the photocurrent density for photoanodes at 1.23 V vs. RHE and for photocathodes at 0 V vs. RHE as benchmark performance metrics. However, these metrics by
themselves are irrelevant for a practical device, since the operational potential, as
outlined in the previous sections, will never be at 0 V or 1.23 V vs. RHE and only
show the half-cell potential of a given working electrode and neglecting the (over)potentials used to drive the counter electrode and ionic conductivity losses in the
solution. Therefore, normalized metrics are required to establish a benchmarking
for the performance of different materials in order to make fair comparisons
between materials and systems that are made and tested in different labs across
the globe.
4.4.1
Solar-to-Hydrogen (STH) Conversion Efficiency
Perhaps the most significant metric for measuring the performance and efficiency of
a solar fuel device is the solar-to-hydrogen conversion efficiency (STH). This
efficiency directly relates the input energy (solar irradiation) to output energy
190
W.A. Smith
(electric/chemical energy via hydrogen evolution minus the input-applied potential) via the following equation:
STH

H2 mol=s m2 Gfo, H2 kJ=mol
Pout Pelectrical
Pin
Plight
Plight W=m2
AM 1:5G
4:7
where the numerator contains the output in terms of the rate of gas evolved, H2
(mol H2/s m2) times the Gibbs free energy of formation for hydrogen
(Gof,H2 237 kJ/mol), divided by the total solar irradiation input in terms of the
power density of the incident illumination (Plight in W/m2, or more commonly for
PEC devices, mW/cm2). This expression only holds true when the illumination
source is the direct (or simulated) solar irradiation-matched spectra equal to air
mass global (AM) 1.5. Furthermore, it is only possible to use this equation to
measure the STH of a solar-driven water-splitting reaction when it is possible to
directly measure H2 accurately as a function of time, most importantly for particlebased photocatalysts. When this is not available, for example, it is possible to
convert this equation to a different form that can use a modified version:

STH
jsc mA=cm2 V redox f

Plight mW=cm2

4:8
AM 1:5G
where the numerator now has the power output in terms of the measured current
density jsc in mA/cm2 times the effective potential required to run the desired
reaction (the redox potential of interest, Vredox, which here is the potential converted
from the previously used G 237 kJ/mol 1.23 V), times the faradaic efficiency
of the hydrogen evolution reaction, f. The denominator does not need to have a
term to include the illuminated area of the electrode, since the numerator has the
current density in terms of current per unit area already included.
It is important to note that the STH is measured in a 2-electrode configuration,
and all the potentials applied must be taken between the working and counter
electrode, i.e. it is not possible to use a 3-electrode system and use the potential
applied to a working electrode against a reference electrode.
While the focus of this chapter and the discussion is on the solar to hydrogen
conversion efficiency of the solar water-splitting reaction, a similar metric can be
applied to general solar fuel systems, where hydrogen is not the reduction product
via water splitting, but where, for example, the reduction of CO2 to different
chemical fuels is achieved. In such a case, it is straightforward to calculate the
solar to fuel conversion efficiency, SFE, by the following equation:

SFE
jOP mA=cm2 V redox f

Ptotal mW=cm2

4:9
AM 1:5G
where Jop is the operational current density that is directed towards a specific
product. The potential is correlated to the thermodynamic potential for a different
191
fuel-forming reaction, Vredox. This metric is much more difficult to extract from the
current densities, as it is likely that many products are formed during
electroreduction of CO2, and therefore the faradaic efficiency and partial current
density towards a particular chemical reaction are needed, which is very complicated from a practical perspective and is thus not discussed further in this chapter.
4.4.2
Applied Bias Photon to Current Conversion

Efficiency (ABPE)
An additional tool to determine how a photoelectrode is able to convert photons into

usable electrons via a chemical reaction is to observe how the photon to current
conversion efficiency changes with an applied bias using the so-called applied bias
photon to current conversion efficiency (ABPE). This technique is an obvious
extension to the STH efficiency, with the notable difference that this technique
uses an applied bias between the working electrode and counter electrode, while the
STH is measured without the application of any external bias potential. Therefore,
the ABPE can be written as follows;

ABPE

jsc mA=cm2 V redox V app f
Plight mW=cm2
AM 1:5G
4:10
where Vapp is the applied potential between the working and counter electrode. The
utility of using the ABPE measurement is that it uses extra potential to drive the
water-splitting reaction for a given photoelectrode, which may be useful for estimating how a particular photoanode or photocathode may operate in a tandem
device where an extra potential can be supplied by a second photoelectrode or a
photovoltaic cell connected in series. This allows the measurement of a single
component of a tandem device to be used to estimate the overall photocurrent
density and efficiency that could be drawn if it is used in a tandem absorbing device.
The practical aspects of a tandem absorbing device are briefly discussed in
Sect. 4.2.2 of this chapter, and in more detail in Chap. 12.
4.4.3
Spectral Response Measurements
To measure overall performance and conversion efficiencies of a photoelectrode, it

is necessary to use the entire solar spectrum to excite photogenerated charge
carriers in a photoelectrode. However, it is also useful to understand where these
photogenerated charge carriers come from during photoexcitation, i.e. to be able to
tell which photons produce a certain amount of electrons. Therefore, making
photocurrent measurements as a function of individual wavelengths of light is
necessary. Such a measurement can be accomplished with a light source, a monochromator, and a potentiostat.
192
4.4.3.1
W.A. Smith
Incident Photon to Current Conversion Efficiency (IPCE)
While STH remains the single most important figure of merit to measure the
performance of a PEC material/device, other techniques can be used to provide
essential information of how the material/device works. These metrics are essential
to assess the origin of how a material performs, so that its practical limits can be
defined, and hopefully then overcome with optimized engineering. One such
diagnostic technique is the incident photon to current conversion efficiency
(IPCE), which may also be referred to as the external quantum efficiency (EQE).
The IPCE/EQE measures the efficiency of converting an individual photon to an
extractable electron via the following formula:
electron flux mol=s
photon flux 2mol=s

j mA=cm hcVm
ph
P mW=cm2 nm
IPCE IPCE EQE
4:11
where jph is the photocurrent density, h is Planks constant, c is the speed of light,
(therefore hc can be simplified to 1239.8 Vm), P is the power of light at a particular
wavelength, and is the wavelength of irradiation. To make accurate IPCE measurements, a light source, monochromator, and potentiostat are required in order to
have a spectral distribution that is selective by wavelength, while at the same time
the current density generated at each wavelength needs to be measured. In addition,
it is required that such a measurement takes place in a 3-electrode configuration, so
that the potential of the working electrode can be varied and measured against a
reference electrode. This is in sharp contrast to the measurement configuration
needed for obtaining the STH, which is most important for defining the overall
efficiency of a material, while measuring IPCE is more of a diagnostic tool to tell
more detailed information about an electrode and to help determine the performance limiting factors.
The technique of obtaining IPCE is very useful and relevant for PEC materials
characterization, but has its limitations for what it can tell about the total efficiency
of a system. For example, it is assumed that for the output of the IPCE measurements, i.e. the electron flux, 100 % is used for the evolution of hydrogen and oxygen
and not for a side or back-reaction. Therefore, it is necessary to couple IPCE
measurements with H2 and O2 quantification to ensure that the water oxidation/
reduction reactions being driven by the individual photons show faradaic efficiency,
and thus all the converted photons are only consumed in the water-splitting
reaction. A typical IPCE plot for a BiVO4 photoanode and an a-SiC photocathode
illuminated from the front-side and back-side are shown in Fig. 4.17a, b,
respectively.
Interestingly, the IPCE can be used to estimate the maximum obtainable photocurrent under AM 1.5 irradiation by the following relationship:
193
Fig. 4.17 IPCE data for (a) a BiVO4 photoanode illuminated from the front and backside, and
held at a potential of 1.23 V vs. RHE, and (b) a a-SiC photocathode illuminated from the front and
backside, and held at a potential of 0 V vs. RHE
J AM 1:5 IPCE ed
4:12
where JAM 1.5 is the total photocurrent density under solar irradiation (mA/cm2),
is the photon flux of the solar irradiation (photons/(m2s)), and e is the elementary
charge (C). While this is not a direct or 100 % accurate way to estimate the
photocurrent density of a material under AM 1.5 solar irradiation, it can give a
close estimate if a solar simulator is not available in a particular laboratory, and
only IPCE testing equipment is available. A correlation between IPCE (integrated
photocurrent) and information provided by JV measurements is essential for
ensuring consistency of measurements.
4.4.3.2
Absorbed Photon to Current Conversion Efficiency (APCE)
The IPCE measures the total amount of electrons converted from all of the incident
photons (broken down into individual wavelengths), and thus is useful to estimate
the maximum possible current that can be extracted by a photon source. However,
this technique inherently takes into account all of the photons that are incident on a
photoelectrode (i.e. light that is either reflected or transmitted through the sample)
and converted to usable (i.e. able to drive the water redox reactions) electrons. This
is certainly not the case for a practical semiconductor material, and therefore it is
also useful to normalize the IPCE by the absorbed spectrum of a sample, which
results in the absorbed photon to current conversion efficiency (APCE), or internal
quantum efficiency (IQE).
194
W.A. Smith
IPCE
APCE APCE IQE
A

j mA=cm2 hcVm
ph
P mW=cm2 nm A
4:13
where A is the absorptance as a function of wavelength. The APCE is primarily

used as a tool to determine the optimal thickness of a material to maximize the
light-absorbing path length through a semiconductor.
4.5
Summary and Conclusions
This chapter serves to introduce the reader to the important aspects of measuring the
performance and efficiency of photoelectrochemical water-splitting materials. In
particular, the considerations for designing a PEC cell are discussed in the context
of the materials used (electrodes, electrolyte, membranes) and the different configurations that photo- and electrocatalysts can be combined to make an overall watersplitting device. In addition, standard measuring equipment and techniques are
summarized to aid the reader in the basic materials used in PEC testing. Finally,
several important efficiency and performance metrics are established to determine
the actual usefulness of the measured data, and how this should be compared to
other samples made in different labs across the world. It is hoped that this chapter
serves as a general introduction to the testing and efficiency definitions for PEC
water splitting so that the following chapters are more accessible and understandable on a fundamental level.
Acknowledgments The author gratefully acknowledges Bartek J. Trzesniewski, Ibadillah
A. Digdaya and Fatwa F. Abdi for assistance with several of the figures, Dr. David Vermaas for
contributions to the membrane section, and the MECS group at TU Delft for helpful discussions.
The author is also very grateful for generous funding from Towards BioSolarCells (grant FOM
03), the NWO VENI scheme, and the CO2-neutral Fuel program of NWO/FOM/Shell (project
APPEL).
References
Abdi FF, Han L, Smets AHM, Zeman M, Dam B, van de Krol R (2013) Efficient solar water
splitting by enhanced charge separation in a bismuth vandate-silicon tandem photoelectrode.
Nat Commun 4:2195
Appleby J, Delahoy AE, Gau SC, Murphy OJ, Bockris JOM (1985) An amorphous silicon-based
one-unit photovoltaic electrolyzer. Energy 10:871
Arai T, Sato S, Kajino T, Morikawa T (2013) Solar CO2 reduction using H2O by a semiconductor/
metal-complex hybrid photocatalyst: enhanced efficiency and demonstration of a wireless
system using SrTiO3 photoanodes. Energ Environ Sci 6:1274
195
Arakawa H, Shiraishi C, Tatemoto M, Kishida H, Usui D, Suma A, Takamisawa A, Yamaguchi T

(2007) Solar hydrogen production by tandem cell system composed of metal oxide semiconductor film photoelectrode and dye-sensitized solar cell. Proc SPIE 6650:665003
Bendert RM, Corrigan DA (1989) Effect of coprecipitated metal ions on the electrochemistry of
nickel hydroxide thin films: cyclic voltammetry in 1M KOH. J Electrochem Soc 136:1369
Bornoz P, Abdi FF, Tilley SD, Dam B, van de Krol R, Graetzel M, Sivula K (2014) A bismuth
vanadate-cuprous oxide tandem cell for overall solar water splitting. J Phys Chem C 118:16959
Brillet J, Yum J-H, Cornuz M, Hisatomi T, Solarska R, Augustynski J, Graetzel M, Sivula K
(2012) Highly efficient water splitting by a dual-absorber tandem cell. Nat Photonics 6:824
Chae KJ, Choi M, Ajayi FF, Park W, Chang IS, Kim IS (2007) Mass transport through a proton
exchange membrane (nafion) in microbial fuel cells. Energy Fuel 22:169
Chen S, Wang L-W (2012) Thermodynamic oxidation and reduction potentials of photocatalytic
semiconductors in aqueous solution. Chem Mater 24:3659
Conell RS, Corrigan DA, Powell BR (1992) The electrochromic properties of sputtered nickel
oxide films. Sol Energy Mater Sol Cells 25:301
Corrigan DA, Knight SL (1989) Electrochemical and spectroscopic evidence on the participation
of quadrivalent nickel in the nickel in the nickel hydroxide redox reaction. J Electrochem Soc
136:613
Cox CR, Lee JZ, Nocera DG, Buonassisi T (2014) Ten percent solar-to-fuel conversion with
nonprecious materials. Proc Natl Acad Sci U S A 111:14057
Digdaya IA, Han L, Buijs TWF, Zeman M, Dam B, Smets AHM, Smith WA (2015) Extracting
large photovoltages from a-SiC photocathodes with an amorphous TiO2 front surface field
layer for solar hydrogen evolution. Energ Environ Sci 8:1585
Fujishima A, Honda K (1972) Electrochemical photolysis of water at a semiconductor electrode.
Nature 238:37
Gaillard N, Chang Y, Kaneshiro J, Deangelis A, Miller EL (2010) Status of research on tungsten
oxide-based photoelectrochemical devices at the University of HawaiI. Proc SPIE
7770:77700V
Gillespie MI, van der Merwe F, Kriek RJ (2015) Performance evaluation of a membraneless
divergent electrode-flow-through (DEFT) alkaline electrolyser based on optimisation of electrolytic flow and electrode gap. J Power Sources 293:228
Han L, Abdi FF, van de Krol R, Liu R, Huang Z, Lewrenz H-J, Dam B, Zeman M, Smets AHM
(2014) Efficient water-splitting device based on a bismuth vanadate photoanode and thin-film
silicon solar cells. ChemSusChem 7:2832
Hashemi SMH, Modestino MA, Psaltis D (2015) A membrane-less electrolyzer for hydrogen
production across the pH scale. Energ Environ Sci 8:2003
Haussener S, Xiang C, Spurgeon JM, Ardo S, Lewis NS, Weber AZ (2012) Modeling, simulation,
and design criteria for photoelectrochemical water-splitting systems. Energ Environ Sci 5:9922
Hernandez-Pagan EA, Vargas-Barbosa NM, Wang T, Zhao Y, Smotkin ES, Mallouk TE (2012)
Resistance and polarization losses in aqueous buffer-membrane electrolytes for water-splitting
photoelectrochemical cells. Energ Environ Sci 5:7582
Hickner MA, Ghassemi H, Kim YS, Einsla BR, McGrath JE (2004) Alternative polymer systems
for proton exchange membranes (PEMs). Chem Rev 104:4587
Hickner MA, Herring AM, Coughlin EB (2013) Anion exchange membranes: current status and
moving forward. J Poly Sci B: Polymer Phys 51:1727
Jacobsson TJ, Fjallstrom V, Edoff M, Edvinsson T (2015) A theoretical analysis of optical
absorption limits and performance of tandem devices and series interconnected architectures
for solar hydrogen production. Sol Energy Mater So Cells 138:86
James BD, Baum GN, Perez J, Baum KN (2009) Technoeconomic analysis of photoelectrochemical (PEC) hydrogen production. Directed Technologies Inc., Arlington, VA, US DOE
Contract no GS-10F-009J
Kainthla RC, Zelenay B, Bockris JOM (1987) Significant efficiency increase in self-driven
photoelectrochemical cell for water photoelectrolysis. J Electrochem Soc 134:841
196
W.A. Smith
Khaselev O, Turner J (1998) A monolithic photovoltaic-photoelectrochemical device for hydrogen

production via water splitting. Science 280:425
Khaselev O, Bansal A, Turner J (2001) High-efficiency integrated multijunction photovoltaic/
electrolysis systems for hydrogen production. Int J Hydrogen Energ 26:127
Klahr B, Gimenez S, Fabregat-Santiago F, Hamann TW, Bisquert J (2012) Water oxidation at
hematite photoelectrodes: The role of surface states. J Am Chem Soc 134:4294
Koper MTM (2011) Thermodynamic theory of multi-electron transfer reactions: Implications for
electrocatalysis. J Electroanal Chem 660:254
Kuhl KP, Cave ER, Abram DN, Jaramillo TF (2012) New insights into the electrochemical
reduction of carbon dioxide on metallic copper surfaces. Energ Environ Sci 5:7050
Lewis NS, Nocera DG (2006) Powering the planet: Chemical challenges in solar energy utilization. Proc Natl Acad Sci U S A 103:15792
Licht S, Wang B, Mukerji S, Soga T, Umeno M, Tributsch H (2001) Over 18% solar energy
conversion to generation of hydrogen fuel; theory and experiment for efficient solar water
splitting. Int J Hydrogen Energ 26:653
Lin GH, Kapur M, Kainthla RC, Bockris JOM (1989) One step method to produce hydrogen by a
triple stack amorphous silicon solar cell. Appl Phys Lett 55:386
Luo J, Im J-H, Mayer MT, Schreier M, Nazeeruddin MK, Park N-G, Tilley SD, Fan HJ, Gratzel M
(2014) Water photolysis at 12.3% efficiency via perovskite photovoltaics and earth-abundant
catalysts. Science 345:1593
McCrory CCL, Jung S, Peters JC, Jaramillo TF (2013) Benchmarking heterogeneous
electrocatalysts for the oxygen evolution reaction. J Am Chem Soc 135:16977
McCrory CCL, Jung S, Ferrer IM, Chatman SM, Peters JC, Jaramillo TF (2015) Benchmarking
hydrogen evolving reaction and oxygen evolving reaction electrocatalysts for solar water
splitting devices. J Am Chem Soc 137:4347
McDonald MB, Ardo S, Lewis NS, Freund MS (2014) Use of bipolar membranes for maintaining
steady-state pH gradients in membrane-supported, solar-driven water splitting. ChemSusChem
7:3021
McKone JR, Lewis NS, Gray HB (2014) Will solar-driven water-splitting devices see the light of
day? Chem Mater 26:407
Miller EL, Paluselli D, Marsen B, Rocheleau RE (2005) Development of reactively sputtered
metal oxide films for hydrogen-producing hybrid multijunction photoelectrodes. Sol Energy
Mater Sol Cells 88:131
Mor GK, Varghese OK, Wilke RHT, Sharma S, Shankar K, LaTempa TJ, Choi K-S, Grimes CA
(2008) p-tpye Cu-Ti-O nanotube arrays and their use in self-biased heterojunction photoelectrochemical diodes for hydrogen generation. Nano Lett 8:1906
Morisaki H, Watanabe T, Iwase M, Yazawa K (1976) Photoelectrolysis of water with TiO2covered solar-cell electrodes. Appl Phys Lett 29:338
Nocera DG (2012) The artificial leaf. Acc Chem Res 45:767
Nozik AJ (1976) p-n photoelectrolysis cells. Appl Phys Lett 29:150
Osterloh FE (2013) Inorganic nanostructures for photoelectrochemical and photocatalytic water
splitting. Chem Soc Rev 42:2294
Park JH, Bard AJ (2006) Photoelectrochemical tandem cell with bipolar dye-sensitized electrodes
for vectorial electron transfer for water splitting. Electrochem Solid-State Lett 9, E5
Peckham TJ, Holdcroft S (2010) Structure-morphology-property relationships of nonperfluorinated proton-conducting membranes. Adv Mater 22:4660
Pinaud BA, Benck JD, Seitz LC, Forman AJ, Chen Z, Deutsch TG, James BD, Baum KV, Baum
GN, Ardo S, Wang H, Miller E, Jaramillo TF (2013) Technical and economic feasibility of
centralized facilities for solar hydrogen production via photocatalysis and photoelectrochemistry. Energ Environ Sci 6:1983
Reece SY, Hamel JA, Sung K, Jarvi TD, Esswein AJ, Pijpers JJH, Nocera DG (2011) Wireless
solar water splitting using solicon-based semiconductors and earth-abundant catalysts. Science
334:645
197
Rocheleau RE, Miller EL, Misra A (1998) High-efficiency photoelectrochemical hydrogen

production using multijuntion amorphous silicon photoelectrodes. Energy Fuel 12:3
Rossmeisl J, Qu Z-W, Zhu H, Kroes G-J, Norskov JK (2007) Electrolysis of water on oxide
surfaces. J Electroanal Chem 607:83
Roy SC, Varghese OK, Paulose M, Grimes CA (2010) Towards solar fuels: photocatalytic
conversion of carbon dioxide to hydrocarbons. ACS Nano 4:1259
Sakai Y, Sugahara S, Matsumura M, Nakato Y, Tsubomura H (1988) Photoelectrochemical water
splitting by tandem type and heterojunction amorphous silicon electrodes. Can J Chem
66:1853
Sathre R, Scown CD, Morrow WR III, Stevens JC, Sharp ID, Ager JW III, Walczak K, Houle FA,
Greenblatt JB (2014) Life-cycle net energy assessment of large-scale hydrogen production via
photoelectrochemical water splitting. Energ Environ Sci 7:3264
Sato S, Arai T, Morikawa T, Uemura K, Suzuki TM, Tanaka H, Kajino T (2011) Selective CO2
conversion to formate conjugated with H2O oxidation utilizing semiconductor/complex hybrid
photocatalysts. J Am Chem Soc 133:15240
Seger B, Castelli IE, Vesborg PCK, Jacobsen KW, Hansen O, Chorlendorff I (2014) 2-Photon
tandem davice for water splitting: comparing photocathode first versus photoanode first
designs. Energ Environ Sci 7:2397
Simons R (1993) Preparation of a high performance bipolar membrane. J Membrane Sci 78:13
Trotochaud L, Mills TJ, Boettcher SW (2013) An optocatalytic model for semiconductor-catalyst
water-splitting photoelectrodes based on in situ optical measurements on operational catalysts.
J Phys Chem Lett 4:931
van de Krol R (2012) in Photoelectrochemical Hydrogen Production ed. by R. van de Krol and
M. Gratzel and. Springer, Berlin, p 73
Varcoe JR, Atanassov P, Dekel DR, Herring AM, Hickner MA, Kohl PA, Kucernak AR, Mustain
WE, Nijmeijer K, Scott K (2014) Anion-exchange membranes in electrochemical energy
systems. Energ Environ Sci 7:3135
Vargas-Barbosa NM, Geise GM, Hickner MA, Mallouk TE (2014) Assessing the utility of bipolar
membranes for use in photoelectrochemical water-splitting cells. ChemSusChem 7:3017
Vermaas DA, Sassenberg M, Smith WA (2015) Photo-assisted water splitting with bipolar
membrane induced pH gradients for practical solar fuel devices. J Mater Chem A 3:19556
19562

Photoelectrochemical Cell Design, Efficiency, Definitions, Standards, and Protocols

Uploaded by

Copyright:

Available Formats

Photoelectrochemical Cell Design, Efficiency, Definitions, Standards, and Protocols

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Photoelectrochemical Cell Design, Efficiency, Definitions, Standards, and Protocols

Uploaded by

Copyright:

Available Formats

Chapter 4

Photoelectrochemical Cell Design, Efficiency,

The storage of solar energy into chemical energy through photoelectrochemical

Eo 1:23 V G 237:22 kJ=mol

E 0:00 V vs: NHE

In an alkaline environment (pH 14), these red-ox equations then become:

Eo 0:401 vs: NHE

2H2 O l 2e ! 2OH- aq H2 g E 0:828 V vs: NHE

4 Photoelectrochemical Cell Design, Efficiency, Definitions, Standards. . .

The Photoelectrochemical Cell

A photoelectrochemical reaction takes place in a photoelectrochemical cell. At the

4 Photoelectrochemical Cell Design, Efficiency, Definitions, Standards. . .

PTFE lid with feedthroughs for counter

Holes for cell

Fused silica window

4 Photoelectrochemical Cell Design, Efficiency, Definitions, Standards. . .

utilization, etc., will not be discussed in this chapter, as it is elaborated on in explicit

4 Photoelectrochemical Cell Design, Efficiency, Definitions, Standards. . .

Carrier lifetime (s)

Diffusion length (nm)

Photovoltaic Cell + Electrocatalyst (PV + EC)

The most straightforward approach to achieve efficient solar-driven water splitting

4 Photoelectrochemical Cell Design, Efficiency, Definitions, Standards. . .

4 Photoelectrochemical Cell Design, Efficiency, Definitions, Standards. . .

Fig. 4.7 A hypothetical

Photovoltaic Cell + Photoelectrode (PV + PEC)

4 Photoelectrochemical Cell Design, Efficiency, Definitions, Standards. . .

Fig. 4.9 A hypothetical

efficiency, low-cost PEC device, as many benchmark device performances have

Dual Photoelectrodes (Photoanode + Photocathode, i.e. PEC)

While a PV-PEC device offers some advantages and disadvantages compared to a

4 Photoelectrochemical Cell Design, Efficiency, Definitions, Standards. . .

4 Photoelectrochemical Cell Design, Efficiency, Definitions, Standards. . .

Ion Exchange Membranes

To avoid mixing of the produced fuel (hydrogen or carbon-based products) with

Proton Exchange Membranes

Anion Exchange Membranes

As an alternative for the proton exchange membranes, anion exchange membranes

4 Photoelectrochemical Cell Design, Efficiency, Definitions, Standards. . .

provides the additional advantage of using different (stable) pH at either electrode.

Simulated Solar Irradiation Measurements

4 Photoelectrochemical Cell Design, Efficiency, Definitions, Standards. . .

a-SiC photocathodes were grown by plasma-enhanced chemical vapour deposition

4 Photoelectrochemical Cell Design, Efficiency, Definitions, Standards. . .

Determining the Flatband Potential

Similar to the Von mentioned in the previous section, an essential component of a

Fig. 4.16 A representative

Evolved Gas Quantification

To ensure faradaic efficiency in photoelectrodes and to measure this in absolute

4 Photoelectrochemical Cell Design, Efficiency, Definitions, Standards. . .

Solar-to-Hydrogen (STH) Conversion Efficiency

jsc mA=cm2 V redox f

jOP mA=cm2 V redox f

4 Photoelectrochemical Cell Design, Efficiency, Definitions, Standards. . .

Applied Bias Photon to Current Conversion

An additional tool to determine how a photoelectrode is able to convert photons into

Spectral Response Measurements

To measure overall performance and conversion efficiencies of a photoelectrode, it

Incident Photon to Current Conversion Efficiency (IPCE)

photon flux 2mol=s