Project Final Report - fch.europa.eu · Project Final Report FCH JU Grant ... device based on...

PECDEMO is a Collaborative Project co-funded by FCH JU under the call SP1-JTI-FCH.2013.2.5.

GA n°: 621252. Start date: April 1st, 2014. Duration: 36 months.

Project Final Report

FCH JU Grant Agreement number: 621252

Project acronym: PECDEMO

Project title:

Photoelectrochemical

Demonstrator Device for Solar

Hydrogen Generation

Funding Scheme: FP7-JTI-CP-FCH

Date of latest version of Annex I against

which the assessment will be made: 07/07/2016

Period covered: From Apr 2014 to Mar 2017

Name, title and organisation of the scientific

representative of the project’s coordinator:

Helmholtz-Zentrum Berlin

Prof. Dr. Roel van de Krol

Tel: +49 30 8062 - 43035

Fax: +49 30 8062 - 42434

E-mail: [email protected]

Project website address: http://pecdemo.epfl.ch/

PECDEMO Final Publishable Report

TABLE OF CONTENTS

1.1. Executive summary ........................................................................................................ 1

1.2. Summary description of the project context and the main objectives .............................. 2

1.2.1. Context ............................................................................................................................ 2

1.2.2. Approach and main objectives ........................................................................................ 2

1.3. Description of the main S & T results/foregrounds ......................................................... 4

1.3.1. Work Package 1 ........................................................................................................... 4

1.3.2. Work Package 2 ............................................................................................................... 8

1.3.3. Work Package 3 ............................................................................................................. 13

1.3.4. Work Package 4 ............................................................................................................. 17

1.3.5. Work Package 5 ............................................................................................................. 22

1.3.6. Work Package 6 ............................................................................................................. 24

1.4. Potential impact (including socio-economic impact and wider societal implications) and

the main dissemination activities and exploitation of results ................................................... 29

1.4.1. Potential impact ............................................................................................................ 29

1.4.2. Dissemination activities ................................................................................................. 30

1.5. Public website and relevant contact details .................................................................. 33


1

1.1. Executive summary

PECDEMO’s main aim was to develop a photoelectrochemical (PEC) water splitting

device based on low-cost and abundant materials that shows a solar-to-hydrogen

(STH) efficiency of 10%, a stability of 1000 hours, and an active area of at least 50 cm2.

PECDEMO has addressed these challenges by focussing its efforts on three metal oxide

photoelectrode materials (Fe2O3, BiVO4, and Cu2O) and by combining them with a

silicon- or perovskite-based photovoltaic (PV) cell in a tandem configuration. To

improve the efficiency and stability of the metal oxides, modifications were made by

doping, application of protection layers, nanostructuring, and surface

functionalization with co-catalysts for hydrogen or oxygen evolution. Fe2O3 is the most

stable material; lab tests showed negligible performance decrease after 1000 h

operation. A new hydrogen treatment method significantly improved the

performance of BiVO4 photoanodes, resulting in a 9.2% STH efficiency for a small-area

dual BiVO4/Fe2O3 photoanode/Si PV tandem cell. PECDEMO’s highest efficiency

achieved for small-area devices was 16.2%, obtained for a Ga2O3/Cu2O nanowire

photocathode coupled to a silicon PV cell using a dichroic mirror for photon

management. The highest large-area photocurrent densities were obtained for Cu2O,

giving an unprecedented 3.5 mA/cm2 for a 50 cm2 photoelectrode.

Various large-area cell designs for were studied, resulting in an optimized design that

features an open path for sunlight from the front to the back window, with counter

electrodes placed at both sides of the central photoelectrode. CFD simulations were

used to ensure an optimal flow path of the electrolyte, resulting in efficient removal of

gas bubbles and good thermal management; the temperature of the cell did not

increase above 55°C even under 17-suns concentrated light. Based on this design, a

modular array of four PEC cells of 50 cm2 each was constructed for field tests on the

SoCRatus facility at DLR in Cologne. The cell design showed limited cross-over of H2,

but the efficiencies for BiVO4 and Fe2O3 were modest under concentrated sunlight –

presumably due to poor carrier transport in these materials.

Two conceptually new innovations were made to further improve the PEC concept. A

power management scheme that allows co-generation of electricity and hydrogen;

in combination with active light management, the PEC efficiencies can exceed those

of PV-electrolyzer systems. The second one is the use of auxiliary NiOOH/Ni(OH)2

electrodes, which avoids the need to separate H2/O2 reaction products within the

same cell. This significantly reduces the overall complexity and costs of the concept.

Plant design studies showed that cooling is a crucial issue, especially under

concentrated sunlight. Life-cycle analyses revealed that the PEC-PV approach is

potentially best in class in terms of global warming potential. Economic analysis

showed that PEC-PV generation can compete with PV-driven electrolysis. However,

STH efficiencies higher than 8%, solar concentration factors > 30, cell temperatures

above 60°C, and active areas approaching 1 m2 should be pursued.

Finally, all PECDEMO targets (10% efficiency, 1000 h stability, 50 cm2) have been

individually achieved, but meeting them simultaneously with a single system remains

a major challenge to be addressed.


2

1.2. Summary description of the project context and the main

objectives

1.2.1. Context

Sunlight is by far the largest sustainable source of energy, and there is little doubt that

it will play a major role in any conceivable future energy scenario. One of the main

challenges for the large-scale use of solar energy is its intermittent nature, which

requires intermediate storage solutions. An attractive pathway to achieve this is by

directly converting an abundant resource, such as water, into hydrogen using sunlight.

The hydrogen can then be used directly as a fuel, or further processed into liquid

hydrocarbons. These ‘solar fuels’ have up to 100 times higher energy and power

densities than the best batteries and can be stored indefinitely.

PECDEMO aimed at developing a PhotoElectroChemical DEMOnstrator that

splits water into hydrogen and oxygen under solar irradiation. By integrating the light

absorption and electrolysis functionalities into a single device, significantly lower

balance-of-systems costs than coupled photovoltaic-electrolysis systems are, in

principle, possible. Efficient and cost-effective solar hydrogen production would thus

solve one of the major challenges for a solar-driven society, i.e., that of efficient large-

scale storage of solar energy. However, before this dream becomes reality, some hard

technological and economic targets have to be met. As outlined in the call that

PECDEMO addressed, solar-to-hydrogen energy conversion efficiencies of 8-10% have

to be achieved and lifetimes of more than 1000 h need to be demonstrated. Only

then will there be a realistic chance to meet the FCH-JU’s cost target of 5 € per kg H2

and can this technology have a significant impact on society.

1.2.2. Approach and main objectives

Building on the breakthroughs achieved in the highly successful EU project

“NanoPEC”, PECDEMO partners aimed to develop a module-sized hybrid tandem

device for solar water splitting based on a stable metal oxide photoelectrode as a

wide-bandgap top absorber and an efficient photovoltaic solar cell as a small-

bandgap bottom absorber. Based on earlier work by the partners, three candidates

were selected as promising metal oxide photoelectrode materials: Fe2O3, BiVO4, and

Cu2O. The stability and durability of the photoelectrodes was planned to be

enhanced through functionalization with efficient electrocatalysts, by applying

selective transport layers and protective coatings, and selection of suitable electrolyte

solutions and operating conditions. The photovoltaic cells were to be optimized for

tandem operation with the metal oxide photoelectrodes. Here, silicon-based

photovoltaic cells and the emerging class of perovskite PV cells have been selected

as the most suitable candidates.

The second aim was to demonstrate the scalability of this technology by

combining multiple devices into a larger water splitting module. Nearly all previous

efforts in the field of photoelectrochemical water splitting have been done on < 1 cm2

cells, with only very few exceptions. At such small length scales, ion transport between


3

the electrodes is sufficiently fast. At larger length scales, however, resistive losses due

to mass transport limitations in the electrolyte quickly start to dominate the overall

performance. Innovative cell designs are needed to minimize these losses and to

manage the transport of photons, electrons, and ions in the water splitting system.

To achieve the project goals, five science and technology objectives were defined:

1. To demonstrate a chemically stable and highly efficient stand-alone hybrid

water splitting cell based on a metal oxide photoelectrode in tandem with a

photovoltaic solar cell

2. To develop deposition technologies that are suited for fabricating components

for large-area hybrid PEC-PV devices

3. To design, construct, and test complete large-area hybrid PEC-PV devices for

water splitting

4. To carry out extensive techno-economic and life-cycle analyses based on the

devices’ demonstrated performance characteristics, and evaluate the

potential for large-scale commercialization

5. To build a prototype module consisting of an array of large area devices and

to test this prototype in the field


4

1.3. Description of the main S & T results/foregrounds

1.3.1. Work Package 1

The first work package (WP1) focuses on the development of an efficient and

chemically stable small-area hybrid PEC-PV water splitting device, by utilizing metal

oxides as the PEC materials. Three metal oxides have been chosen to be our main

focus: iron oxide (Fe2O3), bismuth vanadate (BiVO4) and cuprous oxide (Cu2O).

Activities in the last three years include the development of: (1) small-area metal oxide

photoelectrodes, (2) optically-transparent counter electrodes, (3) PEC-PV without

photon management (1st generation device), (4) effective light management

strategies, and (5) PEC-PV including photon management (2nd generation device).

The summary of these activities and achievements within WP1 is as follows.

Fe2O3 photoanode

Prior to PECDEMO, the champion Fe2O3 photoelectrode utilized a resonant light

trapping structure, by depositing hematite on Ag-coated glass substrates.2 One of the

limitation, however, is the oxidation of Ag during the high-temperature deposition of

hematite, which in turn loses its specularity. In order to overcome this, within PECDEMO,

we reversed the order of deposition of the films, so that the Ag layer is deposited after

hematite. In short, hematite was deposited on a Si substrate, and a Ag layer was

deposited on top of this

hematite film. The film is then

flipped over and attached

to another Si substrate.

Finally, the first Si substrate

was removed by dry

etching, thereby exposing

the hematite layer at the

surface. This process

therefore allows Ag layer to

be deposited at low

temperature and in

vacuum, which is ideal for

specular mirror deposition.

The improved results can be seen in the photographs shown in Figure 1a and b.

In addition, we have also developed heteroepitaxial hematite photoanodes with high

crystalline quality. We first deposited a platinum layer (as a bottom contact and a

reflector) on top of (0001) basal plane sapphire, followed by growth of the hematite

layer with pulsed laser deposition. The in-plane alignment of the film stack is

investigated by azimuthal φ-scans of the off axis peaks as shown in Figure 1c and

verifies the epitaxial growth of the layers. We are still in the process of optimizing the

heteroepitaxial films for photoelectrochemical performance, and have achieved

photocurrents of 1.8 mA/cm2. Noteworthy, the flat-band potential of heteroepitaxial

Ti-doped hematite films was found to be ~0.2 VRHE, considerably lower than reported

values for polycrystalline hematite photoanodes that typical range between 0.4 and

0.6 VRHE.3 This may open up another route to reduce the potential of hematite

photoanodes. Our efforts in heteroepitaxial hematite can be found in our recent

publication.4

Figure 1. Photographs of (a) the champion ultrathin films

hematite photoanode from 2013 and (b) one of the most recent

photoanodes obtained by the film transfer process. (c) Off-axis

φ scans of Al2O3(104), Fe2O3(104), and Pt(200) reflections.


5

BiVO4 photoanode

We have explored the possibility of depositing BiVO4 with various techniques, such as

drop casting and reactive magnetron sputtering, but the performance of our Co-Pi

catalysed, spray-pyrolysed, W-doped BiVO4 photoanode is still superior.5 At the

beginning of PECDEMO, in the efforts of transitioning towards large scale BiVO4, we

have also improved the reproducibility of our spray pyrolysed BiVO4; we have now a

sample-to-sample photocurrent reproducibility of ~5-10%. To improve the

photocurrent further,

we have explored

several other dopant.

Calcium-doped

BiVO4 was first studied

in order to produce a

p-type conductivity,

but the improvement

is limited due to

segregation of

calcium. A successful

treatment is to anneal

BiVO4 in a mild

hydrogen

atmosphere (2.4% H2

in Ar) at relatively low

temperature of 300

C. The onset potential and plateau photocurrent were both improved by hydrogen

treatment, as shown in Figure 2a. An AM1.5 photocurrent of 4 mA/cm2 was achieved

at 1.23 V vs RHE for a hydrogen treated W-doped BiVO4. At this point, we were limited

by the absorption in our thin film BiVO4. The absorption can be simply improved by

increasing the thickness of the film, but unfortunately the carrier diffusion in BiVO4 (<

100 nm) does not allow efficient carrier transport in a thick film. We therefore

implemented a dual photoelectrode approach (see Figure 2b), where an additional

BiVO4 photoelectrode was placed behind the same BiVO4. This simple approach—

which is commonly used in the field of organic PV but not to large extent in PEC water

splitting—resulted in further photocurrent improvement to 4.8 mA/cm2 at 1.23 V vs RHE

and > 5.4 mA/cm2 at 1.7 V vs RHE. This satisfied the deliverable 1.1 (first generation

metal oxides with a photocurrent of at least 5.4 mA/cm2) of PECDEMO. The 20%

photocurrent improvement (i.e., from 4 to 4.8 mA/cm2) can also be replicated by

implementing a photon management strategy: we deposited a distributed Bragg

reflector (DBR) at the back-side of the substrate and therefore removing the necessity

of having dual photoelectrode (deliverable 1.3 — proof of enhanced performance

using a photon management strategy).

Cu2O photocathode

In the case of Cu2O photocathode, we have focused on developing a

transparent Cu2O photocathode toward the ultimate goal of designing an efficient

PEC-PV stacked tandem configuration. For the efficient transparent Cu2O

photocathode, we optimized the thicknesses of Au underlayer and Cu2O flat film (see

Figure 3a for the structure of the photoelectrode), because these are crucial

Figure 2. (a) AM1.5 photocurrent-voltage curves of W-doped BiVO4

(black), hydrogen-treated W-doped BiVO4 (blue), and dual hydrogen-

treated W-doped BiVO4 (red). The respective labels indicate the

photocurrent value at 1.23 V vs RHE. The schematic of dual BiVO4

photoelectrode is shown in (b).


6

parameters affecting the incident light to the PV device placed behind the

photocathode. Consequently, a photocurrent density was reached up to 5.4 mA cm- 2

at 0 V vs RHE (deliverable 1.1 PECDEMO) in pH 5 electrolyte under 1 sun condition using

3 nm Au underlayer and 260 nm Cu2O thickness (Figure 3b). This performance was

comparable to that achieved on standard photoelectrodes on thick Au substrates,

with the advance that the device showed transmittance of around 35 % for

wavelengths longer than 550 nm. We improved the transparency further by

developing an alternative underlayer in the form of Cu-doped NiO. The transmittance

was further increased by ~10-20 % for wavelengths longer than 550 nm, while

maintaining the PEC performance.

The performance of transparent Cu2O photocathode was still low compared

to the theoretical performance. It is mainly attributed to the imbalance of the carrier

diffusion length and the light absorption depth of Cu2O.6,7 Nanostructuring is a

promising way to solve this problem, enabling the further improvement of Cu2O

photocathode performance. We recently succeeded to prepare well characterized

Cu2O nanowire array photocathodes through electrochemical anodization and

thermal annealing. The Cu2O nanowire arrays were high-quality and pure and were

adapted into the photocathode device structure by depositing AZO/TiO2 overlayers

and RuOx catalyst (Figure 3c and d). Remarkably, photocurrent densities up to 8 mA

cm-2 at 0 V vs RHE were reached in pH 5 electrolyte under 1 sun condition, with

Figure 3. (a) False-color cross-section SEM image of a Cu2O-based photocathode, indicating the

different underlayers and overlayers. (b) PEC performance of Cu2O photocathode with different

thicknesses on 3 nm Au underlayer substrates. (c) Scanning electron microscope image of Cu2O

nanowire with AZO/TiO2 overlayers. Inset shows transmission electron image of a single composite of

Cu2O/AZO/TiO2/RuO2 nanowire. (d) Linear sweep voltammetry scan under chopped illumination (1

sun) in the pH 5 electrolyte of Cu2O nanowire photocathode with AZO (black) and Ga2O3 (red)

overlayers.


7

photocurrents exceeding 10 mA cm-2 at more negative potentials. With the

photocurrent level improved, we finally enhanced the performance of Cu2O

photocathode by replacing the AZO overlayer with Ga2O3. Ga2O3 has a minimal

conduction band offset with Cu2O, resulting in a maximized built-in potential and an

anodic shift of the onset potential up to 1.0 V vs RHE, as shown in Figure 3d.

Our improvement efforts on Cu2O photocathodes can be found in our recently

published articles.8-10

PEC and PV stability

The stability of our photoelectrodes have also been investigated under long-term

AM1.5 exposure. Fe2O3 photoanode is known to be highly stable, and we show for the

first time a stability data of up to 1000 hours (See D1.4), achieving our deliverable 1.4

target (less than 10% performance decrease after 100 hours of operation). No

noticeable degradation was observed; full details on this study were recently

published.11 BiVO4 photoanode is expected to be stable in neutral pH electrolyte, but

we observed a photocurrent decrease within the 100-hour measurement period (See

D1.4). The decrease is, however, not related to material degradation, but due to sub-

optimal PEC cell design. Bubbles formed rapidly and trapped at the surface of BiVO4,

causing decreased effective surface area. More optimal cell design is expected to

fully resolve this issue. In alkaline environment, protection layer consisting of TiO2 and

Ni successfully improves the stability, although it is still limited to less than one hour.

Cu2O photocathodes’ stability is shown to be enhanced with the protection layer

strategy that we developed. Although 10% performance decrease is observed within

~55-60 hours, the improvement in stability is unprecedented for Cu2O photocathodes.

For the PV, perovskite solar cell shows increasing efficiency within the first 500 hours of

measurement, with no noticeable change of efficiency afterwards, up to more than

2000 hours of operation. HIT silicon solar cell shows stable short-circuit current and

open-circuit potential within 100 hours, and a slight decrease of fill factor is observed.

Overall, the efficiency decreases only by less than 4%. Detailed description of our

stability tests and results have been published in a public deliverable 1.4 report.

PEC-PV devices

We then combined the photoelectrodes and PV cells developed within

PECDEMO to form a hybrid tandem water splitting device. First, for the first generation

device (i.e., stacked tandem configuration), we have initially simulated the expected

Figure 4. (a) Schematic setup of our 2nd generation PEC-PV tandem configuration, consisting of a

Cu2O photocathode, an IrO2 anode, a HIT solar cell and a dichroic mirror. (b) Chopped AM 1.5 short-

circuit photocurrent density and calculated STH efficiency of the real PEC-PV tandem device. The

green dashed line is the 10% STH efficiency target.


8

solar-to-hydrogen (STH) efficiency from a combination of our photoelectrodes and

multiple multi-junction solar cells. Based on the simulation, we then fabricated a solar

water splitting device consisting of the dual BiVO4 photoelectrode described above

(see Figure 2a) and a 3-HIT series silicon solar cell delivering an STH efficiency of 7.5%

(Figure 5a). To improve the efficiency further, we attempted to address the

fundamental limitation of BiVO4, which is the relatively large bandgap of 2.4 eV. While

our efforts in decreasing the bandgap of BiVO4 is still ongoing, we have simply

extended the light harvesting ability of the dual photoelectrode by combining a BiVO4

and a Fe2O3 photoanode (Figure 5b). As a result, we obtained STH efficiency of 9.2%,

which fulfils the deliverable 1.2 target (1st generation device showing 8% efficiency)

and is already very close to the deliverable 1.5 target (2nd generation device showing

10% efficiency). Detailed results on this dual BiVO4-Fe2O3 photoanode can be found in

our recent publication.1

Finally, for our 2nd generation device, we assembled a hybrid PEC-PV tandem cell

employing a Cu2O–Ga2O3 NW photocathode (see Figure 3d), a HIT PV cell and an

IrO2 anode. For this demonstration, we adopted the tandem configuration with a 600

nm cut-off wavelength dichroic mirror (as shown in Figure 4a). As a result, we

observed operating AM1.5 short-circuit photocurrent of 6.98 mA/cm2 with 2 HIT PV

cell and 10.96 mA/cm2 with 3 HIT PV cell (Figure 4b). The corresponding STH

efficiencies are 10.3 % (2 HIT PV) and 16.2% (3 HIT PV). This result therefore successfully

fulfil our deliverable 1.5 target, as well as the final PECDEMO WP1 target of a device

showing efficiency larger than 10%. Additional information on these PEC-PV devices

can be found in our public deliverable 1.5 report (see PECDEMO website).

1.3.2. Work Package 2 The second work package (WP2) aims to guide the optimization efforts of PEC-PV

tandem cells (WP1) and modules (WP4) by identifying material degradation processes

and efficiency losses; quantifying their effect on the long-term stability and efficiency;

scrutinizing materials compatibility for stable long-term operation with minimal

degradation and efficiency losses; and optimizing the optical and electrical coupling

of the PEC and PV cells. WP2 has five main tasks.

Figure 5. Schematic illustration of a dual (a) BiVO4-BiVO4 and (b) BiVO4-Fe2O3 photoelectrode used in

tandem configuration with a silicon solar cell. STH efficiencies of 7.5% and 9.2%1 have been achieved

with these configurations, respectively.


9

Task 2.1: Develop and implement diagnostic methods to identify materials

degradation processes. This task involved work on Cu2O photocathodes which are

known to be intrinsically unstable in aqueous electrolyte solutions, as well as on BiVO4

photoanodes which, although kinetically stable in neutral aqueous electrolyte

solutions, are unstable in either acid or base solutions. Fe2O3 photoanodes are

thermodynamically stable in base (alkaline) aqueous electrolyte solutions and

typically they do not degrade even after repeated long-term operation in alkaline

solutions. The efforts to stabilize Cu2O photocathodes and BiVO4 photoanodes were

carried out by EPFL and HZB, using passivation overlayers deposited by ALD, as

described in the WP1. Here in WP2 we present our efforts to diagnose the effectiveness

of the passivation overlayers. In order to identify the root cause for Cu2O and BiVO4

photoelectrode degradation, different techniques were used such as:

electrochemical measurements, SEM, TEM, XPS and more. An example of Cu2O

photocathode characterization is presented as Figure 6. The TiO2 overlayer is shown

by the green color in the TEM image.

These characterisation techniques enable us to develop efficient passivation

overlayers, which supress the degradation process in Cu2O and BiVO4 photo-

electrodes and achieve high stability as published recently in Nano Letters9 and

Nature Communication5. Further information on these characterization techniques

can be found in these publications.

Task 2.2: Develop and implement diagnostic methods to identify and quantify

efficiency losses. Within this task, two new diagnostic methods were developed in

order to identify and quantify efficiency losses due to charge separation and

recombination processes. Operando diagnostics of Fe2O3 photoanodes, carried out

at IIT, provides the photocurrent and photovoltage generated by the photoanode, as

presented in Figure 7.

Figure 6. Cu2O photocathode characterisation by SEM (Left), TEM (right) and

electrochemical measurement (inset).


10

This information is used to extract

the maximum power point and the

intrinsic solar to chemical

conversion (ISTC) efficiency of the

photoanode. This information is

important for the design of the

electrical coupling between the

PEC and PV cells (Task 2.4). Further

information about this diagnostics

can be found in the Journal of

Physical Chemistry Letters12.

Photoelectrochemical

spectroscopy combines PEIS

(Photo Electrochemical

Impedance Spectroscopy), IMPS

(Intensity Modulated Photocurrent

Spectroscopy) and IMVS (Intensity

Modulated Photovoltage

Spectroscopy) to provide new

insights on the charge carrier

dynamics involved in the water

photo-oxidation process. This

method was applied at IIT to investigate charge carrier dynamics in Fe2O3

photoanodes and allowed us to quantify the hole current reaching the hematite –

electrolyte interface and the recombination current at interface as presented in

Figure 8. Further information on this method can be found in Physical Chemistry

Chemical Physics13. Similar analysis was carried out for BiVO4 photoanodes at HZB. We

investigated the effect of Co-Pi catalysts on the surface of BiVO4. It has been shown

by many research groups—including HZB—that Co-Pi effectively improves the

performance of BiVO4 photoanode, due to an increased charge injection efficiency.

However, the true nature of the improvement is not clear. Our IMPS study revealed

that the surface recombination rate decreases by more than 1 order of magnitude

upon deposition of Co-Pi on BiVO4 surface. Surprisingly, the charge transfer rate is not

really affected by the introduction of Co-Pi; in fact, it is slightly decreased. This result is

intriguing since it implies that Co-Pi does not function as a “true” catalyst when

deposited on the surface of BiVO4. Instead, it acts as a surface passivation layer,

Figure 7. A typical current density (J) vs. potential (U)

voltammogram of thin film Fe2O3 photoanodes

measured in 1M NaOH aqueous solution in the dark

(dashed) and under AM1.5G solar simulated

illumination (full).

Figure 8. Deconvolution of the water photo-oxidation current (black dots) of Fe2O3 photoanode into

the hole current (red dots) and recombination current (blue dots).


11

reducing the surface recombination. Further information on this study and the

implication can be found in our recent Chemical Science article14.

Task 2.3: Modeling the optical coupling of the PEC and PV cells in PEC-PV tandem cells.

Detailed modeling of the optical

coupling between the PEC and PV

cells in PEC-PV tandem cells was

carried out for the conventional

stacked cells configuration as well as

for advanced configurations that

employ spectral splitting between

the PEC and PV cells in order to

improve the solar to hydrogen

conversion efficiency of the tandem

cell. For example, at HZB we

modeled and tested BiVO4

photoanodes coupled in tandem

with amorphous and nano-crystalline

silicon PV cells. The results are

presented in Figure 9. Further

information can be found in recent

articles published by HZB1,15,16 and

EPFL17.

Several spectral splitting schemes with passive or active light management were

explored in the PECDEMO project, in order to tailor the degrees of freedom of the

tandem system to make it more efficient (but also more complicated). One example

of an active light management design is presented in Figure 10. When the PEC cell is

active, during hydrogen generation, the incident light is splitted between the PEC

(Absorber 1) and the PV (Absorber 2) as presented in the left figure (A). On the other

hand, if only power generation is required the system turns to its “off” state, presented

in the right figure (B). This active light management design allows the system to actively

control the portions of the light that go to the PEC and PV cells in order to tailor

hydrogen and power productions (see Task 2.4 below). This scheme is especially

attractive for tandem cells that co-generate both hydrogen and electrical power as

explained in ACS Energy Letters18.

Figure 9. The EQE curves of the a-Si:H/nc-Si:H solar

cell, the a-Si:H top junction (black), and the nc-Si:H

bottom junction (red). Dashed curves indicate EQE

spectra of the PV junctions under full AM 1.5

simulated solar illumination; solid curves indicate the

EQ

Figure 10. Active light management design in a two-absorber tandem system. A) Absorber 1 (PEC)

is active (“on”). B) Absorber 1 (PEC) is inactive (“off”).


12

Task 2.4: Modeling the electrical coupling of the PEC and PV cells. Modeling and

testing the electrical coupling of the PEC and PV cells started from simple coupling by

series connection of the two cells1,15-17. This scheme leads to a large loss of electric

power generated by the PV cell, which is typically much larger than the power used

by the PEC cell. In order to rectify this loss, we invented a new design that splits the

power generated by the PV cell between the PEC cell and another consumer, i.e.,

co-generation of hydrogen and power. The electrical power can power auxiliary

systems such as cooling,

flow and compression,

or be sold to the grid.

Further optimization can

be achieved using

power convertors that

enable continuous

tracking of the

maximum power point

of the PEC-PV tandem

system. Applying the

power splitting

approach enables to

overcome the

efficiency of PV-

electrolysis systems, as

presented in Figure 11

and explained in ACS

Energy Letters18

Task 2.5: Predictive

modeling of large-area

PEC-PV tandem cells. The series resistance loss due to the resistance of the transparent

electrode that collects the photocurrent from the photoelectrode in the PEC cell

becomes critical in large-area PEC-PV tandem cells. Figure 12 presents an analysis of

the effect of the series resistance on the photocurrent as a function of the size of the

photoelectrode. The analysis was done for a transparent electrode with a sheet

resistance of 15 /square, which is

typical for FTO-coated glass

substrates (TEC15). It clearly shows the

adverse impact of the series

resistance on large area cells. This

effect must be rectified using metallic

grid lines, as implemented in WP4.

Solar plant design. One of the

greatest challenges in large-scale

solar water splitting plants is the

separation of the hydrogen from the

oxygen and the collection and

transport of the hydrogen from

millions of PEC cells distributed in the

solar field to a central hydrogen

distribution facility. This involves an

immense sealing and piping

constructions that puts a heavy

Figure 12. Calculated photocurrent density as a

function of the photoanode radius for a series

resistance of 15 /square.

0 5 10 15 201

1.5

2

2.5

3

3.5

4

Photoanode Radius [cm]

J (

U=

1.5

V v

s R

HE

) [m

A]

Figure 11. Calculated figure of merit (FOM, the ratio between the

total power produced by the PEC-PV tandem system and the power

produced by the PV system alon) as a function of the fraction of the

total power production that goes toward chemical power

generation


13

burden on the hydrogen production economy. To overcome this challenge we

invented a new PEC cell design with separated oxygen and hydrogen cells.

According to this design, the PEC solar cells produce oxygen that is simply discharged

to the atmosphere, whereas the hydrogen is produced elsewhere in another cell. The

ion exchange between the two cells is mediated with a set of auxiliary electrodes

made of NiOOH and Ni(OH)2 that undergo a reversible redox reaction exchanging

OH- ions with the primary electrodes (i.e., the photoanode and cathode) in the oxygen

and hydrogen cells. This enables centralized hydrogen production far away from the

solar field, as explained in Nature Materials.19


Fabrication of large area TCO-coated glass substrates.

In order to fabricate large-area transparent conducting oxide (TCO) electrodes spray

pyrolysis based fluorine-doped tin oxide (FTO) deposition was developed. This process

allows to prepare custom-made conductive substrates with the resistivity and

transmittance directly related to the FTO thickness deposited (Figure 13). It is a flexible

process that allows coating a large variety of substrates as long as the raw material is

able to withstand the high

temperatures. TCOs on up to

35 x 35 cm² were deposited

and characterized including

double-side coating, e.g.

suitable for tandem devices.

Fabrication of large-area

photovoltaic (PV) cells

A mesoscopic

methylammonium lead

iodide perovskite/TiO2

heterojunction solar cell with

low-cost carbon counter

electrode and full screen

printable process was built

on a monolithic design

without any extra organic hole conducting material (Figure 14). These cells were

claimed to be air stable under illumination. Such a perovskite, fully printable

mesoscopic solar cell was deposited on an FTO covered glass. The mesoporous layers

were infiltrated with perovskite by drop-casting from solution through an 8 µm thick

carbon layer printed on top of the ZrO2. The dense TiO2 layer deposited on the FTO

conducting glass prevents the valence band holes from reaching the FTO-covered

front electrode. With a complete set of new screen printable materials: (1) Ti-Nanoxide

Figure 14. Left: SEM cross section of a monolithic perovskite test cell. Right: Perovskite PV Module

Figure 13. Sheet resistance of FTO with various thicknesses.

Insert: Examples of FTO coated glass.


14

BL/SP (50 nm TiO2 based

electron blocking layer); (2) Ti-

Nanoxide T600/SP (600 nm

nano-TiO2 scaffolding layer);

(3) spacer layer: Zr-Nanoxide

ZT/SP (1 µm nano-ZrO2 layer);

(4) highly conducting carbon

layer: Elcocarb B/SP (5 ~10

µm, 10 ~ 20 ohm/sq resistivity) cell efficiencies over 13 % were achieved on lab scale

(1.2 cm2). Monolithic modules (64 cm2 aperture area) with an optimized design to limit

the electrical losses showed an efficiency of 11% (Jsc =16.3 mA/cm2, Voc = 7.1V). The

perovskite deposition process was automated to improve safety, reproducibility and

productivity. By means of a DC-DC converter the appropriate voltage for biasing the

photo-electrode can be obtained. For direct integration in PEC+PV tandem devices

silicon heterojunction (SHJ) modules were developed, fabricated and analyzed, e.g.

upon various illumination condition conditions. Such SHJ based modules offer the

advantages that high PV efficiencies exceeding 20% are possible and the excellent

near-Infrared light absorption and high voltage makes it appropriate for application

as bottom cell in such a tandem configuration with a wide-gap photo-electrode (PE)

in front. In order to match the required voltage to bias the PE two PV cells have been

interconnected (Figure 15). Modules reached efficiencies of 15.7% under one sun

(Figure 16), which allow for STH efficiency of complete PEC-PV tandems devices

exceeding 8%.

Evaluation of deposition techniques for fabrication of counter-electrodes

Spray coating, a well-known method to prepare thin particle layers on substrates (e.g.

varnishing), was adapted for fabrication of counter-electrodes for the

electrochemical hydrogen and the oxygen evaluation reaction (HER/OER). An

automated setup was built with an airbrush moveable in two axis. The developed ink

contains the catalyst, a volatile solvent mixed of water/ethanol and an additional ion-

conductive binder on basis of Nafion®. Different catalysts (iridium dioxide, platinum

on carbon, Evonik OER rare metal free) were deposited in different loadings on a

porous nickel foam (1.6 mm thickness) as substrate. Ex-situ testing was done in a three-

electrode setup similar to the common rotating disk electrode (RDE) technique. For

measurements at high current densities, electrodes were mounted to a motor, so that

the catalyst coated porous substrate can rotate around its own axis. This leads to an

efficient electrolyte flow right through the electrode enhancing the mass transport,

especially the removal of the produced gas. The polarization curves in Figure 17 show

that the spray-coated electrodes are suitable for the electrochemical water splitting.

Figure 15. Cross section of PE + PV tandem device with two

interconnected Silicon heterojunction cells.

Figure 16. Left: 50 cm² PV module consisting of two Si cells. Right: IV date of SHJ cells and modules

Samplearea

(cm²) (%)

Jsc

(mA/cm²)

Voc

(V)

FF

(%)

Vmpp

(V)

Impp

(mA)

Median of 14 cells (reference) 4 21.0 36.5 728 79.2

Best 5x5 cm² cell cut out of wafer 25 16.0 34.1 699 67.0

Best 2-cell module 50 15.7 17.4 1.40 64.7 1.0 15.3

Module with hematite PE 50 10.6 11.7 1.36 66.8 1.0 10.3


15

Beside the known good activities of state-of-the-art catalysts is the activity of a rare

metal free catalyst similar to IrO2.

Large-scale Cu2O photocathode

Large-area, transparent photocathode consisting of electrodeposited Cu2O thin film,

atomic-layer deposited AZO/TiO2 over-layers and photo-deposited RuOx catalysts

were deposited on up to 5 x 10 cm2 area. For enhanced charge collection we applied

metal grids and contacts on the Cu2O photocathode. By optimizing metal grids and

contacts on edges, we demonstrated that sputtered Cu grids and Paste-based Ag

contacts on edges are effective to improve the charge collection on the large-scale

Cu2O photocathode. Especially, Ag contacts on edges improved the resistive PEC

performance, while Cu grids assisted to enhance photocurrent density. We could get

a photocurrent density of 3.7 mA cm-2 at 0V vs RHE, corresponding to a STH efficiency

of 4.6 % using Cu grids and Ag contacts in pH 5 electrolyte under 1 sun illumination

from LED light source (Figure 18 a, c). We finally introduced the Ga2O3 overlayer to the

Figure 18. (a) A transparent large-scale AZO overlayered Cu2O photocathode with Cu grids. (b) A

transparent large-scale Ga2O3 overlayered Cu2O photocathode with an active area of 5 x 10 cm2.

(c) Linear sweep voltammetry scans in the pH 5 electrolyte under 1 sun illumination from LED light

source of large-scale Cu2O photocathode with AZO overlayer (black), AZO overlayer/Cu grids (blue)

and Ga2O3 overlayer (red).

Figure 17. Polarization curves of deposited counter-electrodes with OER-catalysts (left) and HER-

catalyst (right). Electrodes were made by spraying a catalyst ink on a nickel foam as porous

substrate. The measurements were done in a potentiostatic mode vs. an Ag|AgCl 3M KCl

reference in 1M KOH at 25 °C. At each point, the ohmic cell resistance was determined and IR

correction of applied potential was performed. The geometrical area of the deposited electrode

was 1 cm².


16

large-scale Cu2O photocathode instead of AZO overlayer. Consequently, the

optimized Ga2O3 film by atomic layer deposition was homogeneous on the large-

scale, resulting in an anodic shift of the onset potential up to 0.9 V vs RHE.

(Figure 18 b, c).

Large-scale W:BiVO4 Photo-anodes

Large-scale photo-anodes consisting of TEC 15TM FTO substrate with a

electrodeposited Ni grid, a spray pyrolysis deposited thick 1% Tungsten doped BiVO4

(W:BiVO4) absorber (~ 200 nm), and photo-deposited CoPi catalyst were fabricated

on up to 7 x 12 cm2 area, with an active area of 5 x 10 cm2. In order to limit the ohmic

losses, 200 nm thick and 2 mm wide Ni lines were electrodeposited onto treated FTO

substrates spaced 9 mm apart prior W:BiVO4 deposition, and Paste-based Ag contacts

and Cu tape were placed along the edges. As with the large scale Cu2O

photocathodes, the Cu contacts on edges improved the resistive PEC performance,

while Ni grids assisted to enhance photocurrent density. We finally introduced a rapid

annealing step of the W:BiVO4 photoanodes in 2% H2 Ar 98% atmosphere at 320 oC for

10 mins which shifted the photocurrent onset potential from ~0.4 V to 0.3 V vs RHE and

a maximum photo current density of 1.8 mAcm-2 at 1.25 V vs RHE can was achieved

with W:BiVO4 without CiPi in pH 7 1.0 M KPi, 0.5M Na2SO3 (hole scavenger) electrolyte

(Figure 19). For the optimised large area W:BiVO4 photoanodes with CoPi catalyst and

an active area of 5 x 10 cm2 we could achieve a photocurrent density of 1.5 mAcm2

at 1.23 V vs RHE corresponding to a STH efficiency of 1.85 %, using TEC 15 FTO, Ni

gridlines, and Cu contacts in a pH 7, 2.0 M KPi electrolyte, with a 3 electrode setup,

under 1 sun illumination from a quartz tungsten halogen lamp (Figure 19).

Figure 19. W:BiVO4 photoanode deposited on to TEC 15TM FTO with Ni gridlines b) without Ni

gridlines, with the edge coated with Paste-based Ag contacts and Cu tape, protected with Kapton

tape. c) Photocurrent densities against potential vs. RHE, in a 3 electrode configuration (WE, CE and

Ref.) under 1 sun, in hole scavenger electrolyte (1.0 M KPi, 0.5M Na2SO3). Presented is a comparison

of the performance of the W:BiVO4 photo-anodes for a small scale sample (black), a 50 cm2

sample without gridlines (blue), with gridlines (green) and with gridlines + hydrogen annealing (red).


17

Large area integrated PV-PEC devices were fabricated consisting of a 2 x HIT Silicon

module with an area of 5 x 10 cm2 and one large scale CoPi coated W:BiVO4 photo-

anodes as the front window of the electrochemical cell. The integrated PV-PEC device

achieved an initial photocurrent density of 1.20 mAcm-2 at a potential of 1V vs RHE,

corresponding to a STH efficiency of 1.48 % (Figure 20). Although lifetime studies for a

range of samples showed that over 24 hrs the photocurrent density of the integrated

cell decreased and reached a plateau of ~0.6 mAcm-2 at a potential of 1.1 V

corresponding to a STH efficiency of 0.74 %. It is proposed that the loss in photocurrent

is due to a combination of, the degradation and loss of CoPi catalyst and, the partial

degradation W:BiVO4, which visibly remained on the photoanodes after weeks of

testing.


Within PECDEMO, WP4 was focused on the

design, optimization and assessment of a 50

cm2 PEC cell for efficient water splitting and on

evaluating its potential use under

concentrated solar radiation. To accomplish

such goal four main tasks were initially

identified: the development of both “angled”

and “vertical” PEC devices (Tasks 4.1 and 4.2);

their adaptation for use under concentrated

solar radiation conditions (Task 4.3); and, by the

end of the project, the selection of the optimal

device design (Task 4.4).

The key challenge of designing a 50 cm2 PEC

device is to significantly reduce ionic and electronic resistances keeping the cell/unit

price competitive. These resistances depend on the geometry and volume of the cell,

which were optimized to have the lowest ohmic losses. As mentioned, WP4 strategy

relied on the development of two different designs: i) the so-called “angled” PEC cell

comprising an innovative system to separate the evolved gases, with exemption of

the vertical diaphragm and where the photoelectrode may be back or front

illuminated; and ii) the so-called “vertical” PEC cell with a vertical diaphragm, and

where a zero gap distance between electrodes was pursued. Figure 21 displays the

“angled” and “vertical” PEC devices developed within tasks 4.1 and 4.2, respectively.

Figure 21. “Angled” (left) and “vertical”

(right) PEC devices developed within task

4.1 and task 4.2, respectively.

Figure 20. a) The voltage and current overtime for the integrated PV-PEC device under 1 sun whilst

the 2.0 M KPi pH: 7 electrolyte is stirred. d) Photo of the Large scale PV-PEC cell consisting of 2HIT silicon

PV’s, two platinum coated counter electrode and a CoPi coated W:BiVO4 photoanode with an

active area of 50 cm2.


18

“Angled” PEC Cell Design

The developed

“angled” PEC cell took

into account a set of

important requirements in

its design and

construction: i) ionic and

electronic resistances

were minimized to attain

a maximum voltage drop

of 50 mV; ii) the cell

weight and size were

optimized; and iii) its

capability to operate under different working conditions (temperature, irradiances

and tilted positions) assured. Based on previous assumptions, the PEC cell detailed in

Figure 22 was designed and built.

The cell body is made of transparent acrylic assuring resistance to corrosive

electrolytes and a very acceptable range of operating temperatures20. Two metallic

frames (front and back) screwed to the acrylic embodiment against an O-ring assure

the proper sealing of the electrolyte container. The one placed in the front side is

lacquered in black allowing an illumination area of 5 × 10 cm2 through a transparent

glass or synthetic quartz window Figure 22-a4. The interior side of this front window is

spin-coated with a thin film of TiO2, developed at UPorto, which allows the evolved

gases bubbles to easily slip over the cell’s window increasing the amount of light

reaching to the

photoelectrode up to

9 %.21

Another important

feature of this cell is that

the photoelectrode

works simultaneously as

the cell back window,

allowing a narrower

construction. Thus, a

conductive path

between both sides of

the TCO substrate was

developed and

implemented without

changing the dimensions

of the housing area. To fulfill this requirement a conductive silver frame was printed in

both sides of the substrate and in its lateral area. Thus, the back metallic frame that

holds the photoelectrode works as external contact. Such strategy was never reported

before Figure 23. By placing the metallic counter-electrodes side-by-side to the

photoelectrode it is possible to have an open path for the sunlight to reach the PV cell

in the backside of the photoelectrode in a tandem configuration.

The cell is sealed with a transparent acrylic cap screwed on the top. This cap has

three separated compartments, two for hydrogen collection (internally connected)

and one for oxygen collection Figure 22-a3. External tubes can be directly connected

to these cambers. Between the top cap and the transparent body a Teflon®

diaphragm is placed to avoid liquid passage to the gas-collecting chambers due to

Figure 22. “Angled” PEC cell: a) disassembled in 3D project, b) final

device. 1 – acrylic body; 2 – Teflon membrane; 3 – acrylic cap with

separated chambers for oxygen and hydrogen collection; 4 – front

side black metallic frame; 5 – electrolyte inlet; 6 – right side

electrolyte outlet; 7 – one group of acrylic plates.

a) b)

Figure 23. Scaled-up photoelectrode assembly for the “angled” PEC

cell: a) photo of the in-house printed Ag metal frame on the front

and back sides of the TCO glass substrate; b) 5 × 10 cm2 bare

hematite photoelectrode ready to be assembled in the back side of

the “angled” PEC cell.


19

its high hydrophobicity Figure 22-a2. This membrane imposes, however, a pressure

drop to permeate the evolved gases (between ca. 5.5 mbar and 10 mbar) that the

feeding electrolyte has to compensate; this membrane allows the continuous supply

of electrolyte to the cell. The electrolyte inlet is located in the bottom of the cell to

force the upward movement of produced gas bubbles - Figure 22-a5.

“Vertical” PEC Cell Design

The “vertical” PEC cell

design was based in a PEC

device previously developed

by UPorto – the PortoCell.22

Taking advantage of some

important features from the

PortoCell, it comprises a

reservoir holding the

electrolyte wherein the two

electrodes are immersed and

physically separated by a

diaphragm Figure 24. The main

advantage of this cell is the

existence of a vertical

diaphragm that separates the

oxygen and hydrogen

evolution compartments. This

membrane allows placing

both electrodes very close to

each other – near zero gap configuration, minimizing ionic resistances. The dimensions

of the photoelectrode were set at 7.1 × 7.1 cm2 by the PECDEMO consortium in

accordance with the the original proposal of having a 50 cm2 PEC cell.

In this PEC device the photoelectrode also works as cell window taking advantage

of the new strategy implemented in the “angled” PEC cell. Depending on the

diaphragm transparency and the counter-electrode shape different arrangements

could be exploited to maximize the light reaching the photoelectrode and the PV cell

in a tandem configuration as detailed in D4.2 report. Still, this design does not allow

the existence of an open path for the sunlight to reach the PV cell, which is a major

drawback in comparison to the “angled” PEC cell.

Similarly, to the “angled” PEC cell, this cell is made of transparent acrylic and it has

a cap that allows placing a Teflon® diaphragm to collect the evolved gases

separated from the electrolyte. Alternatively, without the Teflon membrane, this cap

allows collecting the evolved gases together with electrolyte. The cell was engineered

considering continuous electrolyte feeding and gas collection in separate chambers.

Concentrated Solar Radiation

Different concentrator concepts for use with PEC-PV devices were developed and

assessed in the scope of Task 4.3. One of the favourite concepts is the Two-axis tracking

Linear Fresnel Reflector realised before in DLR’s test platform SoCRatus (Solar

Concentrator with a Rectangular Flat Focus ).23 Thus, it was decided to employ the

SoCRatus, which provides homogeneously concentrated sunlight with a

concentration ratio of about 17.5, as the solar concentrator for the final prototype. In

order to prepare the development of the final cell design, the “angled” and “vertical”

PEC cells were tested on the SoCRatus. Their general behavior under varying

Figure 24. The “vertical” PEC cell: a) disassembled in 3D

project, b) final prototype. 1 – acrylic body; 2 – diaphragm

holder; 3 – acrylic cap with separated chambers for oxygen

and hydrogen collection; 4 – front and back side black

metallic frames; 5 – Photoactive electrode; 6 – back window;

7 – metallic counter-electrode; 8 – left side electrolyte inlet; 9

- left side electrolyte outlet.

8

9

a) b)

1

2

3

4 4

56 7

10


20

irradiances and tilt angles was assessed with respect to gas mixing, electrolyte and

cell temperature, electrolyte flow stabilization, and gas separation. The experiments

resulted in the identification of cell specific advantages and drawbacks under

practical conditions. The tested devices mounted in the focal plane of the SoCRatus

are shown in Figure 25.

Final Design

The knowledge and experience gained during the initial tasks of WP4 allowed

designing and building the final prototype that best fulfills the targets of the project,

and which is a combination of both cell designs24. A demonstration module

comprising four identical PEC cells of 50 cm2, combining key features of the "angled"

and "vertical" PEC cells and going beyond, is presented in Figure 26. Each cell has an

open path for the

sunlight, from the front

to the back window,

allowing the use of a

tandem PEC/PV

arrangement in which

the PV-cell is placed in

the back of the

photoelectrode. Ray-

tracing simulations

confirmed the

applicability of the

prototype design for

operation with the

SoCRatus concerning

concentration profiles

on the relevant cell

surfaces, i.e. on

photoelectrodes and

PV modules. The

counter-electrodes (CE) are placed side-by-side to the working-electrode (WE), but

physically separated by an anion exchange membrane to avoid gas mixture. The

modular prototype embodies an acrylic skeleton in which the components detailed

in Figure 26 are assembled. In this arrangement, the active area of each

photoelectrode is 5 × 10 cm2 based on the project target ensuring minimum values of

Figure 26. Single PEC cell identical to the 4 units of the sub-module

prototype used in the field tests. 1 – Photoelectrode (back window); 2

– stainless steel frame with an in-built screw for electrical contact; 3 –

platinized-Ti meshes (CE) fixed against the ionic exchange membrane;

4 – screws for electrical contact with the CE; 5 – electrolyte inlets (Ø =

10 mm); 6 – electrolyte outlets (Ø= 3 mm).

Figure 25. The “vertical” PEC cell (left) and the “angled” PEC cell (right) mounted in the focal plane

of the SoCRatus.


21

ionic and electronic resistances. The electrolyte flowpath inside each compartment

of the cell was optimized in a CFD-based simulator to improve heat dissipation and to

allow efficient collection of gases at the top, preventing its accumulation inside the

reservoir. The final and optimized geometry in terms of fluid pattern is presented in

Figure 27.

Two main inlets located at

the bottom of the cell force the

upward movement of

produced gas bubbles – Figure

27. Inside the cell body there

are two inlet manifolds: i) 10

inlets located in the bottom of

the WE compartment (5 close

to the back window and 5 close

to the front window Figure 27-7

and Figure 27-8, respectively);

and ii) 2 inlets located at the

bottom of each CE

compartment ( = 8 mm, Figure

27-5). The electrolyte flow

pattern created by the inlets

located close to both windows

is important to assure the

bubbles detachment. In this

optimized design the top Teflon

membrane was not considered

to avoid the drawbacks

reported on deliverables D4.1,

D4.2 and D4.3. Alternatively,

without the Teflon membrane,

the evolved gases are

collected together with

electrolyte through the four outlets located in the cell cap, two on top of the WE

Figure 28 CFD simulations of the PEC cell: a) Non-optimized

and b) optimized design in terms of temperature profile

[electrolyte: water; total flow rate: 500 ml·min-1; external

temperature: 25 °C under 10-SUN irradiance] and

electrolyte flow distribution [electrolyte: water; total flow

rate: 500 ml·min-1; external temperature: 25 °C].

Figure 27. The optimized design for the electrolyte container of the individual PEC cell that is part of

the modular prototype: a) front, side and tilted views; b) sectional plans considered for the CFD results.

1 – CE compartments; 2 – WE compartment; 3 – main electrolyte inlet ( = 10 mm); 4 – electrolyte

outlets ( = 3 mm); 5 – inlets of the counter-electrode compartments ( = 8 mm); 6 – inlets of the WE

compartment close to the back window ( = 6 mm); 7 – inlets of the WE compartment close to the

front window ( = 4 mm).


22

compartment and one on top of each CE compartment ( = 3 mm, Figure 27-4). A

transparent acrylic plate (2 mm thick) was placed in the middle of the WE

compartment to enhance the fluid flow towards the outlets. This arrangement of inlets

and outlets allowed creating a uniform upward flow with an efficient collection of the

evolved gases. Figure 28 shows the differences between a non-optimized and

optimized design in terms of temperature profile and electrolyte distribution. Following

the same strategy applied to the “angled” and “vertical” PEC cells, photoelectrodes

work simultaneously as front or back windows. In both configurations the

photoelectrode must have an electrical collecting frame at the semiconductor side,

connected to the back of the substrate where the electrical cables are connected or

the PV cell is installed.

The ultimate goal of WP4 was successfully accomplished; a sub-modular prototype

composed by four individual PEC cells, each with an active area of 50 cm2 was

designed, optimized, built and tested under artificial solar conditions and

concentrated solar radiation in field tests.25 Complementary results are presented

hereafter within the framework of WP6, modular prototype and field tests.


The fifth work package addresses design of process, pilot plant and infrastructure

following by inventory analysis and component sizing. Finally, Life-Cycle-Analysis

(LCA), cost estimation for hydrogen production using PEC-PV technology as well as

benchmarking with common H2 production technologies like steam methane

reforming, coal and biomass gasification, wind and PV electrolysis were performed

and evaluated.

Task 5.1. Design of process, pilot plant and infrastructure

The generation of hydrogen with photoelectrochemical-photovoltaic (PEC-PV)

tandem devices via water splitting finally has to be economically viable and

industrially applicable. The PEC-PV system has to be embedded in suitable processes

and plants. Three hydrogen production and application scenarios were considered:

a single home application (SHA), a hydrogen refuelling station (HRS), and an industrial

process (IP). The SHA refers to a decentralised approach of hydrogen production

rated at 1 kg/d @ 6 bar and subsequent use in fuel cells to provide electrical power

needed in small buildings. The HRS offers hydrogen at a nominal production rate of

400 kg/d @ 810 bar to fuel vehicles such as cars and buses, which carry a pressurised

hydrogen tank, whereas the introduced IP features a nominal production rate of

4,000 kg/d @ 20 bar and addresses utilisation of hydrogen as a feedstock for diverse

processes.

Numerous criteria are relevant for the choice of a suitable location for the hydrogen

production plants (weather conditions, politics, terrain, infrastructure, etc.). However,

the most important criterion for the provision of hydrogen at reasonable costs is a high

level of global irradiance. Thus, Seville (Spain) and Negev (Israel) were identified as

promising locations.

Appropriate plant designs for the three scenarios were elaborated, which comprise

beside the PEC-PV system major components such as pumps, heat exchangers,

compressors, tanks, and blowers. The latter component refers to an air cooling system

that controls the temperature of the electrolyte, which flows through the PEC-PV

system, in order to avoid critical temperatures with respect to efficiency and stability.

According to the project targets a solar-to-hydrogen efficiency (STH) of 8% based on

the higher heating value of hydrogen was considered. Collector sizes between 89.1 m2


23

and 378,139 m2 were calculated depending on the location and the scenario. Active

power management (APM)18 allows in-situ generation of excess electricity and was

implemented in the plant models with 5% solar-to-electricity efficiency. Passive cooling

due to convective and radiative heat transfer from the PEC-PV system and the

electrolyte piping to the ambience is relevant and was taken into account. Moderate

concentration ratios up to 30 were considered. The solar plant should be operated at

temperatures as high as reasonable in terms of stability and efficiency, since higher

temperatures clearly decrease the needed active cooling capacity because of

enhanced passive cooling.

Task 5.2. Inventory analysis and component sizing

Mass and energy flows regarding the main plant components were estimated for 60°C

maximum temperature of the PEC-PV system and 8 K temperature increase between

inlet and outlet of the PEC-PV system. Average operation conditions as well as severest

operation conditions concerning ambient temperature and solar input were

considered in order to assess a) representative mean operating modes of the solar

plants and their mean demands with respect to electricity and water and b) the

required maximum operating capacity of the main components of the solar plants. In

both cases negligible influence of wind was assumed.

Cooling is a crucial issue and the implemented blower of the air cooling system

dominates the electricity demand under severest conditions. However, even under

severest conditions and concentrated sunlight APM completely or to a large extend

covers the electricity demand of the entire plant. Since the electricity demand of the

plant increases only moderately for a concentration ratio of 30 compared to 10 or 20

a concentration ratio of 30 was chosen for further investigation related to the HRS and

the IP.

Task 5.3. Life-cycle-analysis (LCA)

To quantify the environmental impact associated with all the stages during the life of

the product, i.e. from the raw material extraction until disposal or recycling (so-called

“cradle-to-gate” cycle), the LCA was performed in accordance with ISO 1404026 using

GaBi 7.0 software27. The focus of LCA was set on the global worming potential (GWP),

which is a measure of the amount of heat trapped by a certain mass of the gas in

question to the amount of heat trapped by a similar mass of carbon dioxide expressed

as kg CO2 eq per kg of produced hydrogen.

It was found that if grid electricity from local sources is used to meet the electricity

demand of plant components, in all analysed scenarios the GWP impact of the PEC-

PV technology is higher than the “best in class” wind electrolysis technology (1.0 kg

CO2 eq kg-1 H2). However, if no external electricity is needed due to APM, the GWP

impact of PEC-PV technology can be lowered up to 1.4 kg CO2 eq kg-1 H2 assuming

implementation of 1 m2 PEC-PV cells. Moreover, if solar concentration (C = 30) is

combined with APM a new state of the art technology with lowest reported to date

GWP impact of 0.4 kg CO2 eq kg-1 H2 could be obtained.

Task 5.4. Cost estimation

Economic analysis was performed using H2A Hydrogen Production model (version 3.1)

provided by the US Department of Energy (DoE) on its web page28 and levelised costs

of hydrogen production (LCHP) were estimated for three different hydrogen

production scenarios. The H2A Hydrogen Production model is based on process design

assumptions, which were verified by an international H2A team. Required input

parameters to the H2A models include capital and operating costs, efficiencies of


24

used process, plant life as well as financial parameters such as the type of financing,

discounted cash flow rate and desired internal rate of return.

Using H2A Hydrogen Production model levelised costs of hydrogen production were

estimated. It was found out that H2 production using 1 m2 PEC-PV cells will lead to LCHP

values of 9 € kg-1 for SHA scenario, LCHP of 19-23 € kg-1 for HRS scenario, and LCHP of

16-20 € kg-1 for IP scenario. In the case of HRS and IP scenarios higher LCHP values

correspond to the case when electricity from the grid is used and no solar

concentration is applied, while lower LCHP values correspond to the case when no

electricity from the grid is used due to implementation of APM and solar concentration

with C = 30 is used. Higher LCHP values for HRS scenario are caused by high electrical

energy consumption due to compression of H2 to 810 bars, while comparably low

LCHP value for SHA scenario is mainly caused by rather simple hardware required. The

specified conditions of hydrogen at the outlet of the plant have a great influence on

the LCHP.

An increase of STH from 8% to 12% or even 15% would have a significant influence on

the LCHP value. Additionally, selling of generated by PV module excessive electrical

energy would generate a revenue and would lead to further decrease of the LCHP

value.

Task 5.5. Benchmarking

Hydrogen production via PEC-PV water splitting was assessed in the context of

alternative hydrogen production technologies. Steam reforming of methane, which

uses as coal gasification fossil feedstocks and therefore inherently involves the

generation of carbon dioxide, is the dominating hydrogen production technology

today. Prominent technologies which use renewable feedstocks are biomass

gasification and electrolysis powered by electricity produced by wind turbines or PV

modules. Respective global warming potentials and H2 production costs29-35 were

analysed and compared.

PEC-PV water splitting could potentially reach lowest reported to date GWP impact

of 0.4 kg CO2 eq kg-1 H2 followed by wind electrolysis (1.0 kg CO2 eq kg-1 H2) and PV

electrolysis (2.5 kg CO2 eq kg-1 H2), biomass gasification (8.0 kg CO2 eq kg-1 H2), steam

methane reforming (14.5 kg CO2 eq kg-1 H2), and coal gasification (23.7 kg CO2 eq

kg- 1 H2).

Comparison of H2 production costs showed that the hydrogen production costs for all

three considered scenarios (9-23 € kg-1 H2) are higher than estimated costs for steam

reforming of methane (0.8-3.0 € kg-1 H2), coal (0.9-2.1 € kg-1 H2) and biomass gasification

(1.0-4.3 € kg-1 H2), as well as wind electrolysis (4.2-6.4 € kg-1 H2). In the case of PV

electrolysis, which shows most similarity to PEC-PV water splitting since it uses the same

feedstocks, a rather broad range of 5.3-22.4 € kg-1 of costs has been estimated, that

shows a wide overlap with cost ranges determined here for PEC-PV hydrogen

production technology. Though, solar concentration with C = 30, active power

management, and large active area per PEC-PV device are promising approaches

to reduce hydrogen production costs and should be pursued, further efforts have to

be made to reach economic viability. Cost figures could effectively be enhanced by

higher STH maintained at higher operating temperatures (aiming at superseding the

active cooling system) and higher concentration ratios.


WP6, Modular Prototype and Field Tests, as a demonstration work package was

focused on the final assessment of the sub-modular PEC-PV prototype, which


25

embraces four identical 50 cm2 compartments. The optimized cell design with respect

to optical path and electrolyte flow characteristics was developed in the scope of

WP4 as presented above. The performance of the prototype under practical

conditions was evaluated in particular regarding efficiency and stability. The final

demonstration phase began in Nov 2016 meeting MS7 – Start of field tests of prototype

module.25

Two sets of experiments were simultaneously conducted to test the optimized

device design: i) with the 1 x 4 demonstration module array under concentrated solar

radiation at DLR and ii) with an individualized 50 cm2 cell, identical to the four cells

comprised in the sub-modular prototype, under non-concentrated artificial sunlight at

UPorto. Each set of experiments was divided into two campaigns. Bare hematite

photoelectrodes produced by UPorto were used in the first campaign, whereas

bismuth vanadate (BiVO4) photoelectrodes with cobalt oxide/phosphate (CoPi)

catalyst prepared at HZB within the framework of WP1 were installed in the second

campaign. Hematite was the semiconductor selected for the first campaign due to its

high stability under continuous operation.11 HIT silicon mini modules manufactured by

HZB/PVcomB, connected in series to the photoelectrodes, delivered bias voltage to

promote the water splitting reactions.

Tests under Concentrated Sunlight

The experiments under concentrated sunlight were conducted employing DLR’s test

facility SoCRatus23 in Cologne. The prototype was mounted in the rectangular focus

of the two-axis tracking solar concentrator and provided with homogeneous, about

17.5-fold concentrated sunlight. The developed PEC-PV device was implemented in

the set-up using two fluid cycles of the SoCRatus. They both fed the inlets of the

prototype, where the flow was distributed to the hydrogen and oxygen chambers

connected to Fluid Cycle 1 and Fluid Cycle 2 respectively.

The first experimental campaign was carried out with front illuminated hematite

photoelectrodes, whereas in the second campaign two BiVO4 photoelectrodes

equipped with grid lines were installed in each compartment – the first one being back

illuminated as part of the front window, the second one being front illuminated as part

of the back window. HIT silicon mini modules with an active area of 50 cm2 each were

placed behind the back windows realizing a true tandem configuration and delivered

bias voltage to the system. In case of hematite partly an additional bias of +325 mV

Figure 29. Modular prototype equipped with BiVO4 photoelectrodes irradiated with concentrated

sunlight in the focal plane of the SoCRatus with reflective shields to protect sensitive parts of the

setup.


26

was applied to reach a total bias of about 1.6 V. The prototype operating with BiVO4

photoelectrodes under concentrated sunlight in the focal plane of the SoCRatus can

be seen in Figure 29. The total irradiation on the prototype, the achieved average

current density, and the estimated molar flow of generated hydrogen relative to

respective mean values for Campaign 1 and 2 are shown in Figure 30. The hydrogen

flow generally follows the current density with a certain delay due to mixing and

saturation effects in the fluid cycle. With hematite a total experimental time of about

15 h was reached, thereof more than 8.5 h without additional bias and close to 6.5 h

with 325 mV additional bias. A total irradiance of 12.4 kW m-2 on average and of

14.0 kW m-2 in peak time was applied. Current densities of about 0.2 mA cm-2 and

0.5 mA cm-2 as well as maximum hydrogen flows of 924 µmol h-1 and 2,078 µmol h-1

were achieved without and with the additional bias respectively. Efficient product gas

separation was obtained. The daily solar-to-hydrogen efficiency (STH) based on the

higher heating value of hydrogen reached 0.059% with additional bias. Within the

duration of operation the prototype featured stable performance.

Campaign 2 with BiVO4 photoelectrodes covered about 48 h. An applied total

irradiance of 7.85 kW m-2 on average and of 16.5 kW m-2 in peak time was estimated.

The mean current density was calculated to 0.87 mA cm-2 while a maximum value of

1.88 mA cm-2 was obtained on Day 2 at about 13 kW m-2. Hydrogen was produced at

rates up to 6,741 µmol h-1. The daily STH reached 0.42% on Day 5 at comparably low

levels of irradiance. Though a certain degradation of the photoelectrochemical

system could be observed within the duration of the campaign, even after 48 h

operation under demanding conditions the BiVO4 system was still active. Since in both

campaigns a non-proportional dependency of hydrogen formation on the solar input

became apparent, further efforts have to be made in order to allow efficient

exploitation of concentrated sunlight.

a) b)

Figure 30. Total irradiation on the prototype (smoothed ± 30 s), average current density (smoothed

± 30 s), and hydrogen flow relative to respective mean values as well as average solar-to-hydrogen

efficiencies (STH) of the particular days associated with a) Campaign 1 (hematite photoelectrodes,

1 M KOH, 25 °C, 1.9 l min-1, membrane: Fumasep® FAA-3-PK-130) and b) Campaign 2 (BiVO4

photoelectrodes, 0.5 M K2SO4 + 0.1 M K2HPO4/KH2PO4, 30 °C, 1.7 l min-1, membrane: Fumasep® FAA-3-

PK-130 + Nafion® NE-1110 / Nafion® NE-1110 / none / none).


27

Tests under Non-Concentrated Sunlight

To assess the stability of the optimized hybrid PEC-PV device under non-concentrated

sunlight an experimental

setup was assembled at

UPorto applying the sulphur

plasma lamp system AS 1300

V 2.0 (Plasma International

GmbH) which provides

1,000 W m-2. In this setup the

PEC cell operates at a

constant bias potential

provided by a HIT Si-PV

module. The medium term

test was performed in two

experimental campaigns: i)

with bare hematite

photoelectrodes and ii) with

BiVO4 photoelectrodes. A

constant electrolyte feeding

of ca. 200 ml min-1 was

promoted using a peristaltic pump and a water bath was used to keep the electrolyte

temperature constant under operation.

Again, hematite was the first photoelectrode tested due to its high stability under

continuous operation; in this experiment the water bath was set to operate at ca.

45 °C.20 At the initial instant a photocurrent density of 0.62 mA cm-2 was produced by

the hematite photoelectrode at 1.6 V in a 2-electrode configuration. Accordingly, to

supply the hematite photoelectrode with the necessary bias potential of ca. 1.6 V, two

50 cm2 PV modules were connected in series and an Autolab potentiostat was used

for continuous monitoring the photocurrent produced by the semiconductor over

1,000 h – Figure 31.

The presented results show that the PEC-PV device remained stable over 1,000 h

(approximately 42 days) delivering an average photocurrent density of ca.

0.43 mA cm-2. The photocurrent oscillations along the polarization curve are due to

periodic interruptions to obtain the J-V curves; a slight decrease on the photocurrent

Figure 31. Polarization curve of the hematite photoelectrode

obtained under a constant bias of 1.6 V and simulated solar

irradiance. Enlarged view of the polarization between 96 h and

192 h.

a) b) Figure 32. a) J-V characteristics of the 50 cm2 hematite photoanode prepared by spray pyrolysis,

before starting the stability test, 0 h ( ), and after 1,005 h ( ) under simulated sunlight; b) J-V

characteristic curves for the two 50 cm2 Si-PV modules connected in series before starting the

stability test, 0 h ( ), and after 1,005 h ( ) under simulated sunlight.

0.0

0.2

0.4

0.6

0.8

1.0

0.8 1.3 1.8

Ph

oto

cu

rre

nt D

en

sit

y /

mA

·cm

2

Applied Potential / V

0 h

1005 h

0

2

4

6

8

10

12

14

16

0.0 0.5 1.0 1.5 2.0 2.5 3.0

Ph

oto

cu

rre

nt D

en

sit

y /

mA

·cm

2

Applied Potential / V

0 h

1005 h


28

density before stabilization was always observed at the initial instants of each testing

period. The latter, can be easily seen in the zoom out of the polarization curve in Figure

31. This initial photocurrent decay, before stabilization, should be related to the

generation of long-lived holes during the formation of a space charge layer, which

can oxidize water on a timescale of around 1 s.36 Figure 32 plots the J-V curves

obtained at the initial instant and at the end of the stability test for the two

components that comprise the hybrid PEC PV device. Both components, the hematite

photoelectrode and the PVs modules, presented very similar performances before

and after 1,000 h of operation with no evidence of corrosion or degradation.

BiVO4 photoelectrodes were tested at UPorto in a second experimental campaign; in

this case the electrolyte was kept at ca. 25 °C. Considering the characteristic

performance of BiVO4 (Figure 33-a) a single 50 cm2 PV module was enough to provide

the necessary bias to the PEC cell for promoting water electrolysis. In this test the setup

operated at the interception point of the J-V curves of BiVO4 photoelectrode and HIT

PV module – Figure 33-a). The photocurrent history over 24 h is plotted in Figure 33-b).

During this period an average photocurrent density of 0.41 mA cm-2 was recorded at

1.28 V, corresponding to a STH efficiency of 0.61%. Over this time the BiVO4

continuously decreased with a current density loss rate of 2 nA cm-2 s-1.

Similar to the test with hematite, the individual performance of the BiVO4

photoelectrode and the PV module was assessed at the beginning and at the end of

the test; the performance of the PV modules remained unchanged after 24 h of

operation. However, from Figure 33-c it can be extracted that the performance of the

BiVO4 photoelectrode continuously decreased after operating 24 h. The latter may be

explained by the material detachment observed during the stability test.

a) b) c)

Figure 33. a) J-V characteristics curves: 50 cm2 BiVO4 photoelectrode (• ) under simulated sunlight and in 0.1

M KPi, obtained in a 2-electrode configuration; single 50 cm2 Si-PV HIT module (• ) under simulated sunlight;

b) Polarization curve of the BiVO4 photoelectrode obtained in a 2-electrode configuration under simulated

sunlight; c) J-V characteristics curves of 50 cm2 BiVO4 photoelectrode, before starting the stability test, 0 h

( ), and after operating 24 h ( ) under simulated sunlight and in 0.1 M KPi, obtained in a 3-electrode

configuration.


29

1.4. Potential impact (including socio-economic impact and

wider societal implications) and the main dissemination

activities and exploitation of results

1.4.1. Potential impact

By achieving its main project goals, PECDEMO has made an important step forward in

the development of efficient, stable, and scalable water splitting concepts. The small-

area solar-to-hydrogen efficiencies1 of up to 9.2% (BiVO4/Fe2O3/Si-HIT) and even 16.2%

(Cu2O/3-HIT) are amongst the highest ever reported for this concept, and have put

Europe at the forefront of efforts in this field. Moreover, PECDEMO has demonstrated

the very first large-area (50 cm2) metal oxide-based PEC-PV water splitting systems that

are based on a true tandem design, i.e., with a wide-bandgap absorber in front of a

smaller-bandgap PV cell. These activities have attracted the interest of Toyota; as a

direct result of the PECDEMO project, one of the project partners (HZB) has recently

started a small seed project with (and funded by) Toyota to further explore

photoelectrochemical water splitting devices.

Although the project represents a significant step forward, we are still far away from a

viable PEC-based technology for solar water splitting. Specifically, the efficiencies for

the large-area devices are still modest. Moreover, fulfilling all three requirements

(efficiency, stability, and scalability) within a single system remains a major challenge.

Nevertheless, with our scaling work we pushed the limit for real application and

performed important pioneering work to reveal (and overcome) limitation

mechanisms and paved the way for solutions, which are of great importance for future

work and coming projects towards commercial PEC-PV applicability.

On the systems level, cooling turned out to be an important aspect that has received

little attention in the field. While all these technical issues can be addressed, the

inherent complexity of the overall process tends to drive up the costs, and makes it

challenging to compete with alternative approaches that make use of mature

technologies, such as PV-driven electrolysis. While this can be partly remedied by

developing more efficient materials, especially light absorbers, innovative new

concepts may be needed in order to achieve the necessary breakthroughs.

PECDEMO has proposed several innovative solutions that may help achieve these

breakthroughs. Examples are the PEC-PV power management strategy (i.e., co-

generation of electricity and hydrogen) and the auxiliary electrode concept. These

concepts have been published in high-ranking journals and are likely to have a

1 STH efficiencies in PECDEMO are based on the enthalpy of hydrogen, 286 kJ/mol (1.48 eV)


30

significant impact on future efforts in the field. Continued efforts by multi-disciplinary

teams consisting of materials scientists, chemical engineers, plant designers, and

business developers are needed to further develop photoelectrochemical water

splitting into a viable technology that has a substantial impact on society.

1.4.2. Dissemination activities

Dissemination activities concentrated on four tasks:

- To effectively communicate PECDEMO’s innovative research

- To establish and maintain a web database to foster communication within the

consortium

- To organize two international workshops

- To conduct outreach activities

For Task 1 the following list compiles some relevant publications from PECDEMO

J. Luo et al. (2014), Water photolysis at 12.3% efficiency via perovskite

photovoltaics and Earth-abundant catalysts, Science Vol. 345/Issue 6204,

26/09/2014 1593-1596

J.H. Kim et al. (2016), Hetero-type dual photoanodes for unbiased solar water

splitting with extended light harvesting, Nature Communications Vol. 7 Nature

Publishing Group, 14/12/2016 13380

Landman et al. (2017), Photoelectrochemical water splitting in separate

oxygen and hydrogen cells, Nature Materials N/A Nature Publishing Group,

13/03/2017

J- Luo et al. (2016), Cu 2 O Nanowire Photocathodes for Efficient and Durable

Solar Water Splitting, Nano Letters Vol. 16/Issue 3, American Chemical Society,

09/03/2016 1848-1857

M-K Son et al. (2017), A copper nickel mixed oxide hole selective layer for Au-

free transparent cuprous oxide photocathodes, Energy and Environmental

Science Vol. 10/Issue 4, Royal Society of Chemistry, 01/01/2017 912-918

Tin oxide as stable protective layer for composite cuprous oxide water-splitting

photocathodes

J. Azevedo et al. (2016), Nano Energy Vol. 24, Elsevier Netherlands 01/06/2016,

10-16

P. Dias et al. (2015), Transparent Cuprous Oxide Photocathode Enabling a

Stacked Tandem Cell for Unbiased Water Splitting, Advanced Energy

Materials Vol. 5/Issue 24, Wiley 01/12/2015

J- Luo et al. (2015), Targeting Ideal Dual-Absorber Tandem Water Splitting

Using Perovskite Photo voltaics and CuIn x Ga 1- x Se 2 Photocathodes,

Advanced Energy Materials Vol. 5/Issue 24, Wiley 01/12/2015

J. Luo et al. (2015), Solution Transformation of Cu 2 O into CuInS 2 for Solar

Water Splitting, Nano Letters Vol. 15/Issue 2, American Chemical Society,

11/02/2015 1395-1402

L. Steier et al. (2015), Low-Temperature Atomic Layer Deposition of Crystalline

and Photoactive Ultrathin Hematite Films for Solar Water Splitting, ACS Nano

Vol. 9/Issue 12, American Chemical Society, 22/12/2015 11775-117 83

J. Azevedo et al. (2014), On the stability enhancement of cuprous oxide water

splitting photocathodes by low temperature steam annealing, Energy and

Environmental Science Vol. 7/Issue 12, Royal Society of Chemistry, 01/01/2014

4044-4052


31

G. Morales-Guio et al. (2015), An Optically Transparent Iron Nickel Oxide

Catalyst for Solar Water Splitting, Journal of the American Chemical Society

Vol. 137/Issue 31, American Chemical Society, 12/08/2015, 9927-9936

J. Luo et al. (2016), Bipolar Membrane-Assisted Solar Water Splitting in Optimal

pH, Advanced Energy Materials Vol. 6/Issue 13, Wiley, 01/07/2016

P. Dias, Extremely stable bare hematite photoanode for solar water splitting,

Nano Energy Vol. 23 Elsevier, 01/05/2016 70-79

J. D. Costa et al. (2016), The effect of electrolyte re-utilization in the growth

rate and morphology of TiO2 nanotubes, Materials Letters Vol. 171 Elsevier,

01/05/2016 224-227

Rothschild et al. (2017), Beating the Efficiency of Photovoltaics Powered

Electrolysis with Tandem Cell Photoelectrolysis, ACS Energy Letters Vol. 2/Issue

1, American Chemical Society, 13/01/2017 45-51

G. Segev et al. (2016), High Solar Flux Concentration Water Splitting with

Hematite (#-Fe 2 O 3 ) Photoanodes, Advanced Energy Materials Vol. 6/Issue

1, Wiley 01/01/2016

H. Dotan On the Solar to Hydrogen Conversion Efficiency of Photoelectrodes

for Water Splitting, Journal of Physical Chemistry Letters Vol. 5/Issue 19,

American Chemical Society, 02/10/2014 3330-3334

S. Kirner et al. (2016), Architectures for scalable integrated photo driven

catalytic devices-A concept study, International Journal of Hydrogen Energy

Vol. 41/Issue 45, Elsevier, 01/12/2016 20823-20831

S. Kirner et al. (2015), Quadruple-junction solar cells and modules based on

amorphous and microcrystalline silicon with high stable efficiencies, Japanese

Journal of Applied Physics Vol. 54/Issue 8S1, Japan Society of Applied Physics,

01/08/2015 08KB03

F. F. Abdi et al. (2014), Plasmonic enhancement of the optical absorption and

catalytic efficiency of BiVO4 photoanodes decorated with Ag@SiO2 core–

shell nanoparticles, Physical Chemistry Chemical Physics Vol. 16/Issue 29, Royal

Society of Chemistry, 01/01/2014 15272

S. Kirner et al. (2016), Wafer Surface Tuning for a-Si:H/µc-Si:H/c-Si Triple Junction

Solar Cells for Application in Water Splitting, Energy Procedia Vol. 102 Elsevier

BV Netherlands 01/12/2016 126-135

Zachäus Photocurrent of BiVO 4 is limited by surface recombination, not

surface catalysis, Chemical Science Vol. 8/Issue 5 Royal Society of Chemistry,

01/01/2017 3712-3719

In addition, PECDEMO was represented at conferences with oral and poster

presentations as listed below (most important)

HZB, Oral presentation to a scientific event, Direct current magnetron sputtering

of photoactive BiVO4: Role of stoichiometry on grain size, structure, carrier

mobility and lifetime, 28/11/2016 MRS Fall 2016,Boston, USA

HZB, Oral presentation to a scientific event, Photoelectrochemical Water

Oxidation of BiVO4 Photoanodes with 50 cm2 Active Area 18/04/2017 MRS

Spring 2017,Phoenix, USA

HZB, Oral presentation to a scientific event, Surface and bulk recombination in

spraydeposited BiVO4, 07/04/2015 MRS Spring 2015, San Francisco, USA

EPFL, Oral presentation to a scientific event, Large Scale Cuprous Oxide

Photocathode toward PEC-PV Tandem Demonstrator for Solar-Driven Water


32

Splitting – From Design to Characterization, 28/11/2016 MRS Fall 2016, Boston,

USA

EPFL, Oral presentation to a scientific event, Using potential-dependent

quantum efficiency measurements to probe device characteristics in

photoelectrodes for solar fuels generation, 01/04/2016 MRS Spring 2016,

Phoenix, USA

UPORTO, Oral presentation to a scientific event, Solar Photoelectrochemical

Hydrogen – Technological Advancements, 02/12/2016, MRS Fall 2016, Boston,

USA

UPORTO, Oral presentation to a scientific event, Up Scaled Photoelectro

chemical Device for Solar Water Splitting-Development and Characterization

of a New Design, 02/12/2016, MRS Fall 2016, Boston, USA

Regarding Task 2, the web domain www.pecdemo.eu was obtained and a

comprehensive project website was built, hosted by EPFL. The site went live on July

15th, 2014. The website features many sections, including “About PECDEMO” (Project

Details, Project Description, Consortium), “Partners”, “Activities” (Meetings,

Deliverables, Demonstrations), and “Dissemination” (Publications, Presentations). The

website was continuously updated with current news and publications from the

project. Pictures of the demonstrator device were published on the website on

November 30th 2016.

For Task 3, the first goal of organizing an international conference was successfully

accomplished by realizing the IPS-20 meeting in Berlin in 2014. The meeting, titled “20th

International Conference on Photochemical Conversion and Storage of Solar Energy”

was organized by HZB and chaired by Prof. Roel van de Krol. The conference was a

great success, attracting over 430 participants from 36 countries and featuring 14

plenary speakers, 19 keynote speakers, and hundreds of contributed talks and posters.

Link: http://www.helmholtz-berlin.de/events/ips20/

The second part of the task was to organize a symposium on solar fuels conversion at

a large international conference. To this end, members of the PECDEMO consortium

have co-organized the “Symposium EC4 – Materials, Devices and Systems for

Sustainable Conversion of Solar Energy to Fuels” at the “2016 Materials Research

Society Fall Meeting” in Boston. The five-day symposium took place November 28 –

December 2, 2016, and featured 21 invited speakers, 73 contributed oral

presentations, and 21 poster presentations. The four co-organizers were Roel van de

Krol (HZB), Avner Rothschild (Technion), Matthew Mayer (EPFL), and Todd Deutsch

(NREL), which were able to recruit symposium support by ACS Energy Letters, ACS

Publications, Helmholtz-Zentrum Berlin für Materialien und Energie, Journal of Physics

D: Applied Physics, IOP Publishing, Nature Energy, and Macmillan Publishers Ltd. The

symposium was well-attended and during the presentation of Harry Atwater, the

meeting room was even filled beyond capacity. Especially the PECDEMO project was

well-represented within the symposium, with 16 oral presentations and 5 posters

contributed by members of the project. For detailed information, see the links:

http://www.mrs.org/fall2016/call-for-papers?Code=EC4 (call for papers)

http://www.mrs.org/fall2016/fall-2016-symposia/?code=EC4 (program)

http://www.pecdemo.eu/

http://www.helmholtz-berlin.de/events/ips20/

http://www.mrs.org/fall2016/call-for-papers?Code=EC4

http://www.mrs.org/fall2016/fall-2016-symposia/?code=EC4


33

Outreach activities were mainly undertaken in the form of teaching. PECDEMO’s main

materials of interest (Fe2O3, BiVO4, and Cu2O) and overall approach were extensively

discussed during the following courses and summer/winter schools:

MSc course on “Photo-Electrochemical Energy Conversion”, taught at the TU

Berlin in the winter semester of 2014, 2015, and 2016

Two-hour seminar on “Solar Fuels and Photocatalysis” as part of the MSc course

on “Modern Developments in X-Ray and Neutron Methods for Science and

Technology“, taught at the Free University of Berlin in 2015, 2016, and 2017

Seminar (1/2 day) taught in August 2015 for students of the German Academy

for Renewable Energy and Environmental Technology

(http://www.germanacademy.net/)

QuantSol Summer School, Hirschegg, Austria (September 2015)

EPFL hosted a one-day research symposium “SwissPEC” on the topic of

photoelectrochemical energy conversion, hosted by EPFL on 11 November

2016 In Lausanne

1.5. Public website and relevant contact details

www.pecdemo.eu

Prof. Roel van de Krol

Helmholtz-Zentrum Berlin für Materialien und Energie

Institute for Solar Fuels (EE-IF)

Hahn-Meitner-Platz 1,

14109 Berlin, Germany

Tel. +49 30 8062 - 43035

Fax: +49 30 8062 - 42434

Mail: [email protected]

http://www.germanacademy.net/

http://www.pecdemo.eu/

mailto:[email protected]


34

References

(1) Kim, J. H.; Jang, J.-W.; Jo, Y. H.; Abdi, F. F.; Lee, Y. H.; van de Krol, R.; Lee, J. S.

Nat. Commun. 2016, 7, 13380.

(2) Dotan, H.; Kfir, O.; Sharlin, E.; Blank, O.; Gross, M.; Dumchin, I.; Ankonina, G.;

Rothschild, A. Nat Mater 2013, 12, 158.

(3) Hankin, A.; Alexander, J.; Kelsall, G. Phys. Chem. Chem. Phys. 2014, 16, 16176.

(4) Grave, D. A.; Dotan, H.; Levy, Y.; Piekner, Y.; Scherrer, B.; Malviya, K. D.;

Rothschild, A. Journal of Materials Chemistry A 2016, 4, 3052.

(5) Abdi, F. F.; Han, L.; Smets, A. H. M.; Zeman, M.; Dam, B.; van de Krol, R. Nat.

Commun. 2013, 4, 2195.

(6) Paracchino, A.; Brauer, J. C.; Moser, J. E.; Thimsen, E.; Graetzel, M. J. Phys.

Chem. C 2012, 116, 7341.

(7) Musselman, K. P.; Marin, A.; Schmidt‐Mende, L.; MacManus‐Driscoll, J. L. Adv.

Funct. Mater. 2012, 22, 2202.

(8) Dias, P.; Schreier, M.; Tilley, S. D.; Luo, J.; Azevedo, J.; Andrade, L.; Bi, D.;

Hagfeldt, A.; Mendes, A.; Grätzel, M. Advanced Energy Materials 2015, 5.

(9) Luo, J.; Steier, L.; Son, M.-K.; Schreier, M.; Mayer, M. T.; Grätzel, M. Nano Lett.

2016.

(10) Son, M.-K.; Steier, L.; Schreier, M.; Mayer, M. T.; Luo, J.; Gratzel, M. Energy

Environ. Sci. 2017.

(11) Dias, P.; Vilanova, A.; Lopes, T.; Andrade, L.; Mendes, A. Nano Energy 2016, 23,

70.

(12) Dotan, H.; Mathews, N.; Hisatomi, T.; Gra ̈tzel, M.; Rothschild, A. The Journal of

Physical Chemistry Letters 2014, 5, 3330.

(13) Klotz, D.; Ellis, D. S.; Dotan, H.; Rothschild, A. Phys. Chem. Chem. Phys. 2016, 18,

23438.

(14) Zachaus, C.; Abdi, F. F.; Peter, L. M.; van de Krol, R. Chemical Science 2017.

(15) Han, L.; Abdi, F. F.; van de Krol, R.; Liu, R.; Huang, Z.; Lewerenz, H. J.; Dam, B.;

Zeman, M.; Smets, A. H. ChemSusChem 2014, 7, 2832.

(16) Han, L.; Abdi, F. F.; van de Krol, R.; Dam, B.; Zeman, M.; Smets, A. H. In

Photovoltaic Specialist Conference (PVSC), 2014 IEEE 40th; IEEE: 2014, p 3083.

(17) Luo, J.; Mayer, M. T.; Grätzel, M. In Organic-Inorganic Halide Perovskite

Photovoltaics: From Fundamentals to Device Architectures; Park, N.-G.,

Grätzel, M., Miyasaka, T., Eds.; Springer International Publishing: Cham, 2016, p

285.

(18) Rothschild, A.; Dotan, H. ACS Energy Letters 2017, 2, 45.

(19) Landman, A.; Dotan, H.; Shter, G. E.; Wullenkord, M.; Houaijia, A.; Maljusch, A.;

Grader, G. S.; Rothschild, A. Nat Mater 2017, advance online publication.

(20) Dias, P.; Lopes, T.; Andrade, L.; Mendes, A. J. Power Sources 2014, 272, 567.

(21) Miranda, S.; Vilanova, A.; Lopes, T.; Mendes, A. RSC Adv. 2017, under revision.

(22) Lopes, T.; Dias, P.; Andrade, L.; Mendes, A. Sol. Energy Mater. Sol. Cells 2014,

128, 399.

(23) http://www.dlr.de/sf/en/desktopdefault.aspx/tabid-10961/19338_read-

44914/.

(24) Vilanova, A.; Lopes, T.; Mendes, A. Submitted 2017.

(25) http://pecdemo.epfl.ch/page-113308-en.html.

http://www.dlr.de/sf/en/desktopdefault.aspx/tabid-10961/19338_read-44914/

http://www.dlr.de/sf/en/desktopdefault.aspx/tabid-10961/19338_read-44914/

http://pecdemo.epfl.ch/page-113308-en.html


35

(26) ISO 14040:2006 Environmental management - Life cycle assessment - Principles

and framework. Ed. 2006.

(27) http://www.gabi-software.com.

(28) https://www.hydrogen.energy.gov/h2a_production.html.

(29) Shaner, M. R.; Atwater, H. A.; Lewis, N. S.; McFarland, E. W. Energy Environ. Sci.

2016, 9, 2354.

(30) International Energy Agency, Technology Roadmap Hydrogen and Fuel Cells,

2015.

(31) Ball, M.; Wietschel, M. The Hydrogen Economy: Opportunities and Challenges;

Cambridge University Press, 2009.

(32) Nikolaidis, P.; Poullikkas, A. Renewable and Sustainable Energy Reviews 2017,

67, 597.

(33) Dincer, I.; Acar, C. Int. J. Hydrogen Energy 2015, 40, 11094.

(34) Acar, C.; Dincer, I. Int. J. Hydrogen Energy 2014, 39, 1.

(35) Hosseini, S. E.; Wahid, M. A. Renewable and Sustainable Energy Reviews 2016,

57, 850.

(36) Tachibana, Y.; Vayssieres, L.; Durrant, J. R. Nat Photon 2012, 6, 511.

http://www.gabi-software.com/

http://www.hydrogen.energy.gov/h2a_production.html

Project Final Report - fch.europa.eu · Project Final Report FCH JU Grant ... device based on...

Documents

Transcript of Project Final Report - fch.europa.eu · Project Final Report FCH JU Grant ... device based on...