Outline for Pre-investigation of Electrolysers

Transcription

Outline for Pre-investigation of Electrolysers
PRE-INVESTIGATION
OF
WATER ELECTROLYSIS
PSO-F&U 2006-1-6287
Draft 04-02-2008
Foreword
This report is the result of an investigation of water electrolysis for hydrogen
production in the energy system. The study has been carried out jointly by (1)
Department of Chemistry, Technical University of Denmark (KI/DTU), (2) Fuel Cells
and Solid State Chemistry Department, Risø National Laboratory, Technical
University of Denmark, and (3) DONG Energy. The investigation constituted the
main part of the project 6287 “Pre-investigation of Electrolysis” funded by the Danish
Public Service Obligation programme (PSO) under Energinet.dk.
The overall aim has been to review:
•
•
•
•
•
State-of –the-art for electrolysis technology from small scale to large industrial
scale comprising the following technologies: Alkaline electrolysis (AEC),
polymer electrolysis (PEMEC) and solid oxide electrolysis (SOEC).
The available commercial products of today with respect to performance and
when possible, cost.
The learning curves for the different electrolyzer techniques
The potential for introduction of electrolysis in the Danish utility system in
connection with an extended production of renewable electricity.
The industrial potential in Denmark for in relation to water electrolysis.
The report is structured as follows
After an introduction (chapter 1) and a brief historical part (chapter 2), the general
fundamental properties and characteristics are explained (chapter 3). The components
of the cell, the thermodynamics and the electrical and overall efficiency are treated.
Chapter 4 is a technical review of the specific types of electrolyzers of relevance. This
is a major part of the report and it is based on an extensive literature study. In chapter
5 information and key figures of commercially available electrolyzers are assembled.
In chapter 6 a technical foresight for electrolyzers is attempted based on the review
performed in the previous chapters. Chapter 7 treats the adaptation of water
electrolysis in the Danish energy system.
The extent of the present report ended up somewhat larger that expected by the
authors and most likely not all parts are equally relevant to all readers. However, we
hope and believe that the information collected can be value for a broad range of
stakeholders addressing different aspects of electrolysis ranging from scientific or
technical development to planning of the future energy system. It should be possible
to find the relevant chapters in the report and skip the rest. Moreover, an executive
summary can be found in the beginning.
04. February 2008
KI DTU
Jens Oluf Jensen
Viktor Bandur
Niels J. Bjerrum
Risø DTU
Søren Højgaard Jensen
Sune Ebbesen
Mogens Mogensen
DONG Energy
Niels Tophøj
Lars Yde (HIRC)
2
Summary
Hydrogen is more and more often mentioned as a solution to the tremendous
challenges resulting from the global worming and depletion of oil and gas. Today,
most hydrogen produced worldwide is from fossil fuels, because this has so far been
the most cost efficient. However, hydrogen has to be extracted from water in order to
avoid the pollution problems and resource limitations of the fossil-fuel-based
production technologies.
The rethinking of the energy system away from a fossil based energy system is
ongoing, and the idea of a renewable energy based system, perhaps in combination
with some form of nuclear power is gaining a wider acceptance then previously.
If the challenges pinpointed by the IPCC and others shall be met tremendous
changes in our energy system is mandatory over a limited period of time. Strongly
increased use of renewables requires large scale energy conversion and most likely
techniques for large scale storage to compensate for the production fluctuations
associated with energy sources like wind and sunlight. Moreover, the transport sector
will most likely need a fuel for propulsion for many years, perhaps forever. Fuel can
to some extend be produced from biomass and waste, but these sources will only
partly cover the need.
Conversion from electricity to a fuel is inevitable in a future energy system,
and no matter if this fuel is pure hydrogen or synthetic fuels (methanol, methane,
gasoline etc.) the first step is production of hydrogen from water splitting. Although
there are a number of methods for water splitting, the only realistic technique for this
process at a large scale is called water electrolysis.
Fundamentals of electrolysis
Many different types of electrolysis cells have been proposed and constructed.
The different electrolysis cells can be divided into groups based on the electrolyte.
Table 1-1 presents an overview of the different types of cells. All the cells presented
in Table 1-1 are capable of using H2O as reactant to produce H2. However, only the
solid oxide cell is capable of using CO2 to produce CO.
Table 1-1. Electrolysis cells and their specialities.
Type
Alkaline
Acid
Polymer
electrolyte
Solid
oxide
Charge carrier
OH-
H+
H+
O2-
Reactant
Water
Water
Water
Electrolyte
Sodium or
Potassium
hydroxide
Sulphuric or
Phosphoric
acid
Graphite
with
Pt,
polymer
Water,
CO2
Polymer
Ceramic
Electrodes
Nickel
Temperature
80 oC
150oC
Graphite
Nickel,
with Pt,
ceramics
polymer
80oC
850oC
3
The overall electrolysis reaction (H2O → H2 + ½O2 or CO2 → CO + ½O2) is a
sum of two electrochemical reactions (also called half-cell reactions), which occur at
the electrodes. The electrode where the reduction of reactants or intermediates takes
place is called the cathode. The anode is the electrode where oxidation of reactants or
intermediates takes place.
Both H2O and CO2 electrolysis become increasingly heat consuming with
temperature. Hence at elevated temperatures a significant part of the total energy
demand can be provided as heat according to Figure 1-1. This provides an opportunity
to utilize the Joule heat that is inevitably produced due to the passage of electrical
current through the cell. In this way, the overall electricity consumption and, thereby,
the H2 and/or CO production price can be reduced.
Figure 1-1. Thermodynamics of H2O and CO2 electrolysis at 0.1 MPa.
The thermo-neutral voltage is defined as:
ETn ≡
ΔH f
nF
=
ΔG f
nF
+
T ΔS f
nF
(1)
where ΔH f is the formation enthalpy change, ΔG f is the Gibbs free energy
change, ΔS f is the formation entropy change, T is the temperature in Kelvin, n is the
number of electrons involved in the electrolysis reaction and F is Faradays constant.
Hence, if the cell voltage equals ETn all the produced Joule heat is utilized. If the cell
voltage is above ETn the cell produces surplus heat (waste heat).
If the cell voltage is below ETn, the produced Joule heat does not meet the heat
demand and the cell cools down if heat is not provided by other means.
For both H2O and CO2 electrolysis, ETn at 0.1 MPa, 25 °C is 1.48 V. At 950 °C,
it is 1.29 V and 1.46 V respectively. Hence, electrolysis of a H2O/CO2 mixture at 950
°C can be performed at thermo-neutral conditions at a cell voltage between 1.29 V
and 1.46 V depending on the H2O/CO2 electrolysis ratio.
4
Efficiency
The efficiency, η , of the electrolysis process may be calculated as the higher
heating value (HHV) of one mole of the product divided by the energy consumption,
W, used to produce one mole of the product.
A high efficiency is of course beneficial, however, an economically optimized
production is usually more important. In order to optimize the production economy a
high production rate is necessary. The higher the cell voltage is increased above ε the
higher is the current density and in turn the production rate. When the cell voltage
increases above ETno , surplus heat is produced and the efficiency decreases. Today’s
alkaline cells are typically operated at ~1.9V or higher in order to optimize the
production economy. With SOECs it is possible to achieve economically optimized
production costs and at the same time keep the cell voltage at ETn or slightly above.
Types of electrolyzers
All electrolyzer consists in the simplest principle of two electrode separated by an
electrolyte. So called half cell reaction resulting in the formation hydrogen and
oxygen respectively take place at one electrode each. The role of the electrolyte is to
close the electrical circuit by allowing ions (but not electrons) to move between the
electrodes. Moreover it keeps the produced gasses separated.
Alkaline electrolyzers (AEC) represent a very mature technology that is the current
standard for large-scale electrolysis. The anode and cathode materials in these systems
are typically made of nickel-plated steel and steel respectively. The electrolyte in
these systems is a liquid one based on a highly caustic KOH solution. The ionic
charge carrier is the hydroxyl ion, OH-, and a membrane porous to hydroxyl ions, but
not to H2 and O2 provides gas separation.
Key advantages of this technology include its maturity and its durability.
Key disadvantages are its use of a highly caustic electrolyte and its inability to
produce hydrogen at high pressures. This inability to produce high pressure hydrogen
for storage results in the added need for an external compressor, which adds cost and
complexity to the system.
The cell reactions are:
4H2O(l) + 4e- → 2H2(g) + 4OH-(aq)
4OH(aq) → O2(g) + 2H2O(l) + 4e2H2O(l) → 2H2(g) + O2(g)
(2)
The electrolyte is an alkaline solution in water, typically 30% potassium hydroxide. It
is contained in a porous felt separator traditionally made of asbestos. The electrodes
are made of nickel or nickel plated metal on which a catalyst is applied. The catalyst
can be noble metals like platinum, rhodium or iridium, but a large selection of nonnoble catalysts is also available. The body report reviews extensive information on
materials for electrodes and catalysts. The details are not suited for being summarized
here.
5
Most commercial potassium hydroxide water electrolysers use nickel electrodes and
are operated at 70-80 ºC. Increasing the operating temperature influences the
thermodynamic of the system. Only limited information on electrolysers operated at
elevated temperatures (above 150 ºC) is available.
Increasing the operation temperature for alkaline water electrolysis from the
normal 80°C to above 200 °C may significantly increase the performance and the
electricity to hydrogen efficiency. A possible obstacle for operating at elevated
temperature is the lower stability of the materials. Development of suitable materials
for the cell is necessary in order to develop a large scale water electrolysis plant for
operation at elevated temperatures. At present possibly suitable cell and separator
materials, which are not more expensive than low temperature alkaline electrolyser
materials, have been identified, but the necessary long term (several years) stability
remains to be proven.
Polymer electrolyzer or proton exchange membrane electrolysers (PEMEC) are built
around a proton conductive polymer electrolyte. The electrodes are:
Anode : 2H2O →4H+ + O2 + 4eCathode : 4H+ + 4e- → 2H2
Cell : 2H2O → 2H2 + O2
(3)
Proton exchange membrane (PEM) water electrolysis technology is frequently
presented in the literature as a very interesting alternative to the more conventional
alkaline water electrolysis.
Proton exchange membrane (PEM) water electrolysis systems offer several
advantages over traditional technologies including greater energy efficiency, higher
production rates, and more compact design.
The advantages of the solid-polymer-type cell are that
(a) The electrolyte membrane or diaphragm can be made very thin, allowing
high conductivity without risk of gas crossover, and
(b) the electrolyte is immobilized and cannot be leached out of the cell.
Also it is ecological cleanliness, considerably smaller mass–volume
characteristics and power costs and, that is very important, a high degree of gases
purity, an opportunity of compressed gases obtaining directly in the installation, the
increased level of safety.
The disadvantages of the solid-polymer-electrolyte (SPE) cell are that
(a) the electrolyte costs more than the conventional alkaline solutions and
(b) the electrolyte is corrosive and requires more expensive metal components
to be used in the cell. For these reasons, solid-polymer-electrolyte cells are usually
operated at somewhat higher current densities than cells that use a liquid alkaline
electrolyte.
Normally, different electro catalysts are utilized for the anode (e.g. IrO2) and
cathode (e.g. Pt). When the electrode layers are bonded to membrane, it is known as
the membrane electrode assembly (MEA).
The membrane consists of a solid fluoropolymer which has been chemically
altered in part to contain sulphonic acid groups, SO3H, which easily release their
hydrogen as positively-charged atoms or protons [H+]:
PEM electrolysis technology has fast response time and start-up/shut-down
characteristics. Hydrogen generation starts immediately at ambient conditions.
6
The type of solid oxide electrolyser cell (SOEC) that has been developed, fabricated
and tested at Risø has shown the best steam and CO2 electrolysis performance ever
reported for an electrolysis cell. In general, the main reason for the high performance
of the SOEC technology compared with other electrolysis technologies is the high
temperature (600 - 1000 °C), since both the H2O and CO2 electrolysis reactions
become increasingly endothermic with temperature and the ohmic losses in SOECs
decreases with increasing temperature. More specific, the reason for the high SOEC
performance is that Risø in about two decades has been among the leaders world wide
in developing the close related SOFC technology.
This high performance makes it potentially possible to establish an efficient
H2 and CO production. A practical electricity to fuel efficiency about 90 % seems
realistic. Estimations of the H2 and CO cost, based on the measured performance and
economic assumptions specified, indicate H2 and CO production costs competitive
with today’s crude oil prices. The H2 production cost was found to be 71 US¢/kg
equivalent to 30 $/barrel crude oil using the HHV (higher heating value). The CO
production cost was found to be 5.6 US¢/kg equivalent to 34 $/barrel crude oil using
the HHV. The main part of the production cost was found to be the electricity cost for
the electrolysis operation.
A combined H2 + CO (synthesis gas) production can be catalyzed into various types
of synthetic fuels, such as methanol, DME (dimethylether) and methane. In such a
synthetic fuel production, some reduction in the production price may be achieved by
utilizing the heat from the catalysis reaction for steam generation.
The lifetime of the SOEC is a main issue to be addressed before the
technology is commercially viable. In general, much R&D work is necessary before
this technology is ready for the market.
For the synthesis of organic fuels a carbon source is needed. This source might on the
short term be the flue gasses of the conventional power plants (or optionally oxyfuel
combustion, i.e. combustion of fuels with oxygen from the electrolyzers). On the
longer term, CO2 from the atmosphere might have to be captured. A system for CO2
capture is discussed. It is a closed loop in which Ca(OH)2 absorbs atmospheric CO2
forming CaCO3. Pure concentrated CO2 is later released by heat. The residue, CaO
absorbs H2O forming Ca(OH)2.
Commercially available electrolyzers
The technology and sizes of commercially available electrolyzers vary greatly. In the
present survey of commercial electrolyzers focus is only on technology that can be
useful to the electrical power grid. Consequently, the performance review only covers
alkaline electrolyzers because they are the only ones available for larger scale
hydrogen production.
7
Research - Product development - Commercial
products
Solid Oxide Electrolyses
Proton Exchange electrolyses
Alkaline Electrolyses
Figure 1-2. A graphical indication of the state of development of the different types
of electrolyzers.
6 major suppliers of alkaline electrolysers have been identified as:
•
•
•
•
•
•
Norsk Hydro in Norway
Hydrogenics in Belgium
Iht in Switzerland
AccaGen in Switzerland
Erre Due in Italy represented by H2Indistrial in Denmark
Uralkhimmash in Russia
The plants can be divided in two groups. Atmospheric and pressurized plants. The
atmospheric plant operates at atmospheric pressure of one bar and the pressurized
plant operates at pressures from 4 to 30 bar depending of the make.
Table 1-2. Large capacity electrolyzers from the listed companies.
Western
Supplier
Hydro
Hydrogenics
IHT
AccaGen
Erre Due
Uralkhimmash
Atmospheric
plants
200 to 2000 kW
14 to 1500 kW
1520-3150 kW
Pressure
plants
50 to 300 kW
60 to 240 kW
500 to 3400 kW
7 to 500 kW
100 to 200 kW
20-1250 kW
Electrolysers have the reputation of being very expensive. It is true but often when the
price pr. kW of a specific electrolyser is mentioned the size of the plant is not given.
The specific price of electrolysers (EURO / kW) is strongly dependent of the size of
the plant. The price analysis in Figure 8-1 shows it very clearly. It can be seen that
the price per kW installed capacity vary with a factor of 10 dependent of the size of
the plant.
8
Euro / kW
Prices of Alcaline Electrolysers
10000
9000
8000
7000
6000
5000
4000
3000
2000
1000
0
0
500
1.000
1.500
2.000
2.500
3.000
kW
1 Bar, HYDRO
4 Bar, H2Industrial
16 Bar, HYDRO
6 Bar, AccaGen
10 Bar, Hydrogenics
1 Bar ELT
Figure 1-3. Prices of electrolyzers as a function of production capacity. The figures
are from the present study.
The electricity cost for hydrogen production was calculated on the following basis: 1)
Actual electricity prices through 2006 (varying). 2) The price of the largest
electrolyzer in the study. Assumptions: 1) Electrolyzer depreciation over 10 years,
75% efficiency (HHV). The production cost for hydrogen was then varying between
0.45 and 0.50 kr/kWh for between 25 and 100% use of the electrolyzer (part time
production at lowest electricity prices).
If, moreover, the value of oxygen and heat was included in the calculation the
cost was between 0.35 and 0.40 kr/kWh.
If the electrolyzer efficiency is assumed td to be 100% instead of 75% the cost
is between 0.30 and 0.33 kr/kWh in the same utilisation interval. This limited
reduction is not an argument against research and development of more efficient
electrolyzers, but a very strong indication that there is absolutely no reason to await
more efficient electrolysers to start business development.
The efficiency of an electrolyser is defined as the ratio of the higher heat value (HHV)
of the hydrogen produced and the DC electricity consumption of the electrolyser. The
commercial electrolyzers have an electricity consumption of 4.1 to 4.8 kWh per
normal m3 produced. Using the HHV of 3.5 kWh/m3 hydrogen, the efficiencies can be
calculated to between 85 and 73%.
The following companies were visited during the project:
•
•
•
•
Norsk Hydro, Nottodden, Norway
IHT (former Lurgi), Geneva, Switzerland
Acca-gen, Lugarno, Switzerland
Hydrogenics, Antwerp, Belgium
9
A pilot plant for 125Nm3 H2/hour and a full-scale project for 50000Nm3 H2/hour were
looked for. Issues like efficiency, pressure and technology were discussed with the
companies.
The H2 is meant for methanol production. This process will run at an estimated
pressure of minimum 70 bar, which means that a high-pressure H2 production will be
an advantage.
Technical foresight
Apart from gradual improvement of performance and lifetime as well as cost, two
development lines are discussed. The one is increased working pressure. If hydrogen
is to be stored in pressure tanks or used for fuel synthesis compression work can be
limited if hydrogen is delivered at higher pressure. The theoretical cost in terms of
electrical energy is smaller than what is practically needed for a compressor.
The other development line is increase working temperature. As discussed
above the minimum work (electrical energy) required decreases with temperature, and
this is one of the arguments for SOEC and also to some extent for AEC and PEMEC
at elevated temperatures. Moreover, elevated temperature results in better electrode
kinetics and electrolyte conductivity, both of which lovers the internal losses. If the
excess heat produced is utilized, it is no longer a loss. A possible use of the heat could
be in the district heating system. This would require slightly higher temperatures than
just below 100ºC, which is normal for the present electrolyzers available.
Due to the necessity of water electrolysis in the future energy system, strongly
increase research activity is expected over the years coming. A number of research
groups in Denmark are already getting involved (to some extent a spin-off from fuel
cells) and are thus establishing a good position for electrolysis development in
Demark.
Oxygen can be used in an oxy-fuel combustion process resulting in high temperatures
and highly concentrated CO2 in the flue gas. This CO2 can be useful for the synthesis
of synthetic fuels (methanol, methane etc), as there will be no need for N2 removal.
Synthesis of methanol, methane or other synthetic fuels might be desirable because it
eases storage and later fuelling. It should be studied under which condition the
electrolyzer itself or the system can facilitate this synthesis. Generally, atomic
hydrogen and oxygen (or their respective ions) that appear at the electrolyzer
electrodes are more reactive than the molecular hydrogen and oxygen they form
before they are released as products of the electrolysis process.
In case hydrogen is produced with the aim of storing, say wind energy, the
oxygen produced should be stored as well. The production ratio (1:2) is of course the
same as needed for the back conversion, and fuel cells operated on pore oxygen
(instead of air with only 21% oxygen) performs with a higher efficiency.
Finally, the Danish wind industry might benefit in terms of competitiveness from a
following technology. When wind power is enhanced to cover a larger fraction of the
energy supply in other countries that Denmark and a few more, technology to handle
the produced hydrogen will be asked for.
10
Adaptation of electrolysis in the Danish energy system
The occurrence in recent times of serious blackouts in America and Europe
underscores the fact that additional measures are urgently needed to avoid such costly
incidents. In addition integration of increasing generation capacity from renewable
energy sources is a challenge to the operation of the system. The need and potential
for integrating energy storage or energy conversion in electrical power systems with
high wind penetration is already widely recognised within electric power utilities. In
this context electrolysers - being both flexible electric loads, energy conversion
systems and storage - can increase the flexibility of the system and be an important
measure to allow the integration of additional renewable energy in the Danish power
system.
The individual electrolyser cells require voltage adjusted in the interval 80100% as to change the current from 20% to 100%.
Large-scale electrolyser as e.g. envisaged in DONG Energy’s in the range of several
hundreds of MW’s will have to connect at the transmission level at 132-150 kV.
Connection of large-scale electrolysers in the transmission system can mitigate
transfer capacity problems in the transmission system, which occurs e.g. in periods
with high wind generation. In this way significant costs of transmission line upgrades
can be avoided.
In addition electrolysers may offer possibilities for improving the security of
the system by e.g. including them in remedial action schemes as significant load that
can be disconnected.
Finally, as will be discussed in the following, by applying PWM converter
technology electrolysers may eventually add significantly to the reactive power and
voltage control of the system.
Medium scale electrolysers (50 MW or less) may connect to the distribution system at
voltage levels 50-60 kV or 10-20 kV. Here it is of interest to consider Energinet.dk’s
development plans for a new network structure in Denmark. It’s a concept based on a
two-layer structure where the 150 kV and the 400 kV transmission levels are to be
jointly planned and operated and the local grids below each 150/60 kV transformer
station will constitute network cells in which monitoring and control are performed.
Flexible electrical load (FEL) is defined as variations of consumer load on a shortterm basis as response to price signals. The Danish TSO estimates that the potential
for price flexibility of regular consumers in Denmark is approximately 660 MW. The
integration of electrolysis plants of say 100-200 MW would thus add significantly to
the flexibility. Increasing wind capacity in the system will tend to give larger
fluctuations in electricity prices and increase problems of power overflow when wind
power can cover the complete load. A more flexible demand will mitigate these
problems and give more stable prices, and thus an increased integration of
electrolysers (and other flexible loads, storage and energy conversion systems) would
increase the value of the wind generation.
The choice of electrolyser type has influence on the power control capabilities of the
system. Standard alkaline electrolysers presently in operation are not designed for fast
control. Large electrolysers like Norsk Hydro are designed for continuous operation
and take approx. two hours to start-up and increase production to 100%. Increased
11
cooling can decrease the start-up time. Small alkaline electrolyser cells are available,
which offer 20 seconds start-up time.
A Solid Oxide Electrolyser system can be controlled by regulation of the
temperature (by changing the current through the cells) and at the same time control
the water or steam supply. The thermal control of the cells will be important for the
control of power to the electrolyser. Downwards regulation from 100% to 0% can
happen in approx. 30 seconds, whereas the upward control can take 15 minutes or
more.
PEM electrolysers will offer better performance in terms of control range and
regulation time. A dynamic range of the PEM electrolyser 5–100 % of rated capacity
has been achieved and typically the electrolyser will have a response time from 5-100
% in less than a second.
It is presently required that generation units, which are connected to the transmission
system, are obliged to provide automatic active power reserves for the support of the
power system in case of disturbances. Presently such obligations are not applied to
specific flexible loads. However, in case emergency reserves are not sufficient to
stabilize the power system after the disturbance; frequency protections will
automatically shed loads (i.e. disconnect consumers). Electrolysers may offer such
services and as excellent performance can be achieved in combination with wind
generation.
Finally, with a large fraction of renewables in the energy system water electrolysis is
inevitable even though the technology is not perfect. Efficiencies, short term cost and
the advantages and drawbacks of the different technologies can be, and will be,
discussed. However, a more fundamental question could be “what is the alternative, if
business as usual is not an option?”
12
FOREWORD
2
SUMMARY
3
TABLE OF CONTENT
13
1
INTRODUCTION
15
2
HISTORY OF ELECTROLYZERS
18
3
FUNDAMENTALS OF ELECTROLYSIS
25
3.1
THERMODYNAMICS
3.1.1
Temperature
3.1.2
Thermo-neutral voltage
3.1.3
Pressure
3.2
EFFICIENCY
4
TYPES OF ELECTROLYZERS
4.1
4.1.1
4.1.2
4.1.3
4.2
4.2.1
4.2.2
4.2.3
4.3
4.3.1
4.3.2
4.3.3
4.3.4
4.3.5
4.4
4.4.1
4.4.2
4.4.3
4.4.4
4.5
4.5.2
4.5.3
4.5.4
4.5.5
4.5.6
4.5.7
5
ALKALINE ELECTROLYZERS
The cell
Stacks and systems
State of art
POLYMER ELECTROLYZERS
The cell
Stacks and systems
State of art
HIGH TEMPERATURE ALKALINE ELECTROLYZERS
Thermodynamic of water electrolysis
Gibbs energy
Enthalpy
Materials
Conclusion
SOLID OXIDE ELECTROLYSER CELLS
Introduction
SOEC History and Background
SOEC state of the art at Risø
International SOEC status
ECONOMIC MODELLING OF H2 AND CO PRODUCTION USING SOEC
The experimental results used as input
Economic input
Discussion of the results of the economic calculations
Conclusions on SOEC
Appendix 1. CCASR calculation
Appendix 2. Economy calculation
COMMERCIALLY AVAILABLE ELECTROLYZERS
5.1
PERFORMANCE REVIEW
5.1.1
Western marked
5.1.2
Eastern marked
5.1.3
Prices, Efficiency, CE-Marking, Safety
5.2
COST
5.2.1
Hydrogen as an energy raw material
5.2.2
Electricity costs
5.2.3
Investment costs
5.2.4
Depreciation
5.2.5
Optimum operation of electrolyser plants
5.2.6
Utilization of oxygen and heat.
5.2.7
Excess of wind power
26
26
27
27
28
30
30
30
33
39
62
62
65
66
71
72
72
74
77
78
78
78
79
81
83
84
86
87
94
102
103
104
107
107
110
123
126
129
129
130
133
134
135
136
138
13
5.2.8
An estimate of the European market for electrolysers
5.3
VISITS TO SUPPLIERS OF ELECTROLYSERS
5.3.1
General questions and answers.
5.3.2
Norsk Hydro minutes of meeting.
5.3.3
AccaGen, minutes of meeting
5.3.4
IHT minutes of meeting.
5.3.5
Hydrogenics minutes of meeting.
6
TECHNICAL FORESIGHT
6.1
6.2
6.3
6.4
6.5
7
ADAPTATION OF ELECTROLYSIS IN THE DANISH ENERGY SYSTEM
7.1
7.2
7.2.1
7.2.2
7.2.3
7.2.4
7.2.5
7.3
7.3.1
7.3.2
7.3.3
7.4
7.4.1
7.4.2
7.5
7.6
7.6.1
7.6.2
7.6.3
7.6.4
7.7
7.7.1
7.7.2
8
ELECTROLYSIS IN THE DANISH POWER SYSTEM
THE POWER SUPPLY OF ELECTROLYSERS
The load control
Intermittent operation
AC/DC conversion and power regulating
Converters and converter configurations
Alternative converter configurations
GRID CONNECTION AND INTEGRATION IN THE ELECTRICAL POWER SYSTEM
Large scale electrolysers at transmission levels
Medium and small scale electrolysers at distribution levels
Electrolysers combined with wind generation
FLEXIBLE ELECTRICAL LOAD
The electricity price variations
Electrolysis as flexible load
ANCILLARY SERVICES
ACTIVE POWER CONTROL
Active power control on electrolysers
Primary active power/frequency control
Automatic emergency reserves
Secondary active power control (automatic or manual reserves)
REACTIVE POWER AND VOLTAGE CONTROL
Reactive power control of electrolysers
Reactive power reserves, voltage control and short circuit level
CONCLUSION
8.1
8.2
8.3
9
ALKALINE ELECTROLYZER CELLS (AEC)
POLYMER ELECTROLYZER CELLS (PEMEC)
SOLID OXIDE ELECTROLYZER (SOEC)
REVERSIBLE FUEL CELLS
SYSTEM DEVELOPMENT
ELECTROLYZER TECHNOLOGY
LARGE COMMERCIAL ELECTROLYZERS
FUTURE RESEARCH AND DEVELOPMENT
REFERENCES
138
139
140
145
147
149
154
156
156
157
157
158
159
160
160
160
160
161
161
162
163
164
164
165
166
167
168
169
170
171
171
172
172
172
173
173
175
176
176
177
179
182
14
1 Introduction
It has been a dream for over a century: a zero-emission energy economy in which
power for cars and buildings alike is generated from converting hydrogen in fuel cells
that produce no more than electricity, heat and water. Although hydrogen has been
used as a fuel at various times – for instance as a constituent of town gas in the early
20th century and as a synthetic fuel during wartime, a lot of renewed interest in
hydrogen and fuel cell technology has been generated in recent years. Driven by
diverse concerns such as depletion of oil resources, climate change, urban air
pollution and security of energy supply, rich countries are looking for alternatives to
traditional fuels. Recent advances in technology are also bringing the cost of using
hydrogen down sharply.
The large-scale application of hydrogen technology would involve significant
changes in the energy system. Decisions on whether and how to promote hydrogen
technology are thus strategic choices over different pathways our energy system will
take from today, along with the various environmental, social and economic impacts
they entail.
In spite of the fact that hydrogen is an inherent component of conventional
hydrocarbon fuels, such as oil, natural gas and coal and that it will be used as a fuel
long after these non-renewable energy supplies are gone, the public tends to give
hydrogen little attention, perhaps forgetting or not realizing that it is used in large
quantities every day to fuel cars, heat or cool building, and fertilize grass and crops. It
is produced in enormous quantities as an industrial "intermediate" in the production of
ammonia, fertilizers, methanol and other chemicals and in the refining of petroleum,
but people are unaware of it or its significance in daily life because it maintains a low
profile in combination with other elements as gasoline, diesel, natural gas or fertilizer
rather than as a free substance in its own name. Spurred mainly by the OPEC crisis in
the early '70s, and a growing awareness of the finiteness of natural hydrocarbon
supplies, dedicated researchers around the world have been working diligently to
develop improved and new ways to split water, the only significant renewable source
of hydrogen that is available now and for as long as human life shall survive. The
range of approaches being investigated covers a wide variety of electrolytic, thermal
and photochemical techniques, and while advances are being made on all these fronts,
alkaline water electrolysis has the most significant near-term commercial potential for
recovery of hydrogen from water on a large industrial scale.
Hydrogen is more and more often mentioned as a solution to the tremendous
challenges resulting from the global worming and depletion of oil and gas. However,
hydrogen or subsequent synthetic fuels are only energy carriers, i.e. tools to handle
the energy. An energy amount equivalent to at least the energy content of the
hydrogen (and practically more due to conversion losses) must be supplied by energy
sources like e.g. wind, sunshine, biomass or nuclear. It can always be discussed which
energy sources are primary and which are not, but from a hydrogen energy point of
view, the key thing is that energy from other sources is stored as hydrogen for later
conversion.
Most of the hydrogen produced worldwide today is from fossil fuels, primarily
through steam reforming of natural gas because this has so far been the most cost
15
efficient. Fossil fuels can also be subjected to several other reactions (gasification of
coal, catalytic decomposition, partial oxidation, etc.) to obtain hydrogen [1].
Unfortunately, this fossil-fuel-based hydrogen is not environmentally benign, does not
contribute toward reduction of greenhouse gas emissions [2]. Hydrogen has to be
extracted from water in order to avoid the pollution problems and resource limitations
of the fossil-fuel-based production technologies [3]. This splitting of water can be
achieved through direct electrolysis or via one of the several thermochemical cycles
where the net reaction is the decomposition of water. Thermochemical cycles can in
principle be very efficient [4], but they require that the energy is provided as heat at
very high temperature. Moreover, this high efficiency may not be realized because of
the complexity and poor selectivity of the proposed thermochemical systems. As a
result, the electrolytic decomposition of water, a relatively well-known and
established technology, may possibly be superior to any thermochemical cycle [5, 6].
The production of hydrogen by the electrolysis of water is, in principle, very
simple. The basic electrolysis cell consists of a pair of electrodes immersed in a
conducting electrolyte dissolved in water. A direct current is passed through the cell
from one electrode to the other. Hydrogen is evolved at one electrode, oxygen at the
other, and water is thus consumed from solution. In a continuously operating
electrolysis cell, pure water is continuously supplied, and a continuous stream of
hydrogen and oxygen may be obtained from the electrodes. In practice, electrolysis
cells are more complicated, containing various other components that allow them to
work efficiently and economically. Because the basic electrolysis cell has no moving
parts, it is reliable and trouble-free; and electrolysis represents the least labourintensive method of producing hydrogen.
In addition to the trouble-free operation, electrolysis is the most efficient way of
generating hydrogen under pressure. Increasing the pressure of operation of the cell
results in a higher theoretical voltage requirement to drive the cell, but electrolysis
cells normally work more efficiently at a higher pressure; and the gain in efficiency
usually more than offsets the extra electrical energy required.
An important characteristic of electrolysis is that hydrogen and oxygen are
separated at the same time. This benefit is derived at the expense of having to use a
high "energy form," namely electric power, as the input to the cell.
The rethinking of the energy system away from a fossil based energy system is
ongoing, and the idea of a renewable energy based system, perhaps in combination
with some form of nuclear power is gaining a wider acceptance then previously. This
change is undoubtedly coursed to some extent by the political situation in parts of the
world as well as the limitation of resources, but the ever growing driver today is
apparently the fear of global warming.
If the challenges pinpointed by the IPCC and others shall be met tremendous
changes in our energy system is mandatory over a limited period of time. Strongly
increased use of renewables requires large scale energy conversion and most likely
techniques for large scale storage. There are two reasons for this. 1) The renewable
energy production is typically very fluctuating (e.g. wind, solar) and the production
will not be able to match the demand if a large fraction of the energy supply is
fluctuating like the wind power. 2) The transport sector will for a long time need a
fuel which can be stores onboard vehicles and ships. Battery powered vehicles are
under steady development, by they still have a long way to go especially in terms of
range before they are flexible enough for replacing fuelled vehicles. For heavy
transport like trucks the demands are more severe, and for ships and planes battery
16
systems are even more speculative. Even with a future introduction of a certain fleet
of battery vehicles, the demand for some sort of fuel will be inevitable.
Fuel can to some extend be produced from biomass and waste, but these
sources will only partly cover the need. Some countries plan on depending on nuclear
power, and this way it might be easier to match the overall energy demand, but it
doesn’t solve the problem of the transport sector as the energy still has the form of
electricity or heat, just like the renewables (apart from biomass and waste).
In conclusion, conversion from electricity to a fuel is inevitable in a future energy
system, and no matter if this fuel is pure hydrogen or synthetic fuels (methanol,
methane, gasoline etc.) the first step is production of hydrogen from water splitting.
Although there are a number of methods for water splitting, the only realistic
technique for this process at a large scale is called water electrolysis. In the following
chapters different aspects of water electrolysis are reviewed.
17
2 History of electrolyzers
Luigi Galvani was an Italian physician and physicist who lived and died in
Bologna. Galvani attended Bologna's medicine school and became a medical doctor
just like his father.
In 1783 Galvani dissected a frog at a table where he had been conducting
experiments with static electricity; Galvani's assistant touched an exposed sciatic
nerve of the frog with a metal scalpel, which had picked up a charge. At that moment,
they saw sparks in an electricity machine and the dead frog's leg kick as if in life. The
observation made Galvani the first investigator to appreciate the relationship between
electricity and animation — or life. He is typically credited with the discovery of
bioelectricity.
Galvani coined the term animal electricity to describe whatever it was that
activated the muscles of his specimens. Along with contemporaries, he regarded their
activation as being generated by an electrical fluid that is carried to the muscles by the
nerves. The phenomenon was dubbed "galvanism," after Galvani, on the suggestion of
his peer and sometime intellectual adversary Alessandro Volta. Galvani's report of his
Figure 2-1. Luigi Galvani.
investigations were mentioned specifically by Mary Shelley as part of the summer
reading list leading up to an ad hoc ghost story contest on a rainy day in
Switzerland—and the resultant novel "Frankenstein"—and its electrically reanimated
construct.
18
Figure 2-2. Galvani’s frogs.
Galvani's investigations led shortly to the invention of an early battery, but not by
Galvani, who did not perceive electricity as separable from biology. Galvani did not
see electricity as the essence of life, which he regarded vitalistically. Thus it was
Alessandro Volta who built the first battery, which became known therefore as a
voltaic pile.
Figure 2-3. Alessandro Giuseppe Antonio Anastasio Volta
Alessandro Giuseppe Antonio Anastasio Volta, Conte, was professor of physics at the
University of Pavia from 1779 and became famous for his work in electricity.
Napoleon I made him a count and a senator of the kingdom of Lombardy. Volta
invented the so-called Volta's pile (or voltaic pile); the electrophorus; an electric
condenser; and the voltaic cell.
In most of Galvini's experiments the frogs were mounted on a brass hook and a
muscle spasm was caused when a different metal was used to complete an electric arc
to the brass hook. This was generally a scalpel, or when testing atmospheric
electricity the frog legs were hung against a metal railing during a thunderstorm. The
"metallic electricity" did not explain why the frog leg would occasionally jump
19
without two types of metal on a clear day. But as there was no scientific explanation
for this, Volta concentrated on the two metal theory.
In 1800, he announced a new electrical device, the Voltaic Pile, initially
presented as an "artificial electric organ", in controversy with the claimed autonomy
of animal electricity.
Volta demonstrated that when metals and chemicals come into contact with
each other they produced an electrical current.
This device was made of alternating disks of zinc and copper with each pair
separated by brine soaked cloth. Attaching a wire to either end produces a continuous
current of low intensity. This was the first direct current battery. This put an end (for a
time) to Galvani's theory of animal electricity. It is interesting to note that Volta
described his battery as an electric organ and likened it to the electric organ of the
torpedo fish, which had columnar stacks of cells. The effect was, for the first time in
history, a method for obtaining a continuous electric current. This first electric battery
was described in a letter from Volta to the President of the Royal Society of London,
Sir Joseph Banks, dated 20 March 1800. The letter written in French was read at a
Royal Society meeting on June, 26 and soon published in the Philosophical
Transactions [7]. The untitled original material was given a heading by Banks: “On
the electricity excited by the mere contact of conducting substances of different kinds.
In a letter from Mr. Alexander Volta, F.R.S., Professor of Natural Philosophy in the
University of Pavia to the Rt. Hon. Sir Joseph Banks, Bart. K. B. P. R. S.”.
Figure 2-4. The Voltaic Pile.
On receipt of the letter Banks was also excited. It had to pass through France,
which was then at war with Britain, and Volta seems to have expected problems of
communication. Possibly for that reason he sent his note in two parts. While waiting
for the second part Banks showed the first few pages to Anthony Carlisle, a London
surgeon. He began trying to repeat Volta's experiments immediately. Humphry Davy
said Volta's work was “an alarm bell to experimenters all over Europe” and Carlisle
was the first to prove him right.
In a few weeks Anthony Carlisle, a London surgeon entered a friend, William
Nicholson. Together they replicated Volta's experiments, using Nicholson's doubler to
20
show charges on the upper and lower plates. This meant that they had to connect them
to the electroscope, and it was not easy to maintain a good contact. To overcome this
little problem they added a drop of water to the uppermost disc and inserted the wire
in that. They were surprised to note the appearance of a gas, soon shown to be
hydrogen.
They then took a small tube filled with water from the New River (an artificial
channel completed in 1613 to bring water from Hertfordshire to the City) and inserted
wires from the Voltaic pile at each end. To their astonishment the other suspected
constituent of water, oxygen, did not appear at the same place but at the other wire “at
a distance of almost two inches”. They had discovered electrolysis.
Figure 2-5. William Nicholson (1753 –
1815)
Figure 2-6. Sir Anthony Carlisle (17681840)
Figure 2-7. A simple electrolyzer illustrating the principle.
21
Nicholson and Carlisle used platinum electrodes and separate tubes to collect
the gases evolved at each electrode. Hydrogen gas bubbled from around the cathode
and oxygen gas from around the anode. The ratio was of two volumes of H2 for every
volume of O2.
They discovered that the amount of H2 and O2 set free by the current was
proportional to the amount of current used.
Waiting till the rest of Volta's letter had arrived and been presented to the Royal
Society, Nicholson and Carlisle decided to publish their results, and where better than
in Nicholson's Journal? This humble periodical was the means for conveying to the
world a discovery that led to a new science: electrochemistry. Quickly it published
many other results in this field, including descriptions of the Voltaic pile with DIY
instructions for making one, and early papers on the subject by Humphry Davy.
Not all readers were convinced that actual decomposition had taken place,
especially those who had doubts about the new chemistry. Hydrogen and oxygen
might be compounds of water with (respectively) positive and negative electricity,
they thought. But it did not take long for doubters to be convinced, and Lavoisier's
chemistry received an additional boost. Within a few years electrolysis had been used
by Davy to isolate sodium, potassium, calcium, strontium, barium, magnesium and
lithium. Chemistry would never be the same again.
Year by year, the land was becoming less able to provide enough food for the
increasing population. Concerns grew in Europe, Asia, Australia and America at the
beginning of the 20th century, fuelled by the British chemist William Crookes, who
maintained in his famous speech of 1898 that: England and all civilized nations stand
in deadly peril of not having enough to eat.
Throughout the centuries, natural fertilizers had been the most important means
of increasing crops, but supplies of natural fertilizers depend again on the supply of
animal feed – animals need to eat in order to produce manure.
In his speech, William Crookes indicated where the answer was to be found.
Passing a strong electric current between two poles causes the air to catch fire,
producing nitrous gases, which contain bound nitrogen.
Many people became interested in this question from both a theoretical and an
industrial point of view. An intense technological competition arose, and a number of
patents were taken out in several countries. Two Americans, Bradley and Lovejoy,
together with the company Atmospheric Products Co., developed a method they
believed would be successful at the Niagara Falls in the U.S.
However, although they had access to inexpensive hydroelectric power, the
method didn’t work as planned. Their equipment was damaged after a short time, and
by 1904 they had given up.
There was widespread work in Germany to find a practical solution. In 1903,
Professor Frank revealed that he had produced nitrogen compounds from calcium
carbide. The resulting product, calcium cyanamide, contained around 20 per cent
nitrogen, and could be used as fertilizer.
One of the German companies that started working on arc technology after 1898
was BASF (Badische), under the leadership of the chemist Otto Schönherr and the
electrical engineer Johannes Hessberger. Progress was slow, and sometimes the work
stopped altogether. In the autumn of 1903, Badische was contacted by a Norwegian
engineer, Sam Eyde. This seemed to lead to renewed efforts on a broad front to find
the best technology for extracting nitrogen from the air. The articles “Explosive
winter days in 1903” and “A project of calibre” illustrate the connection. The
22
inquiries by Eyde were no coincidence. There was someone in Norway – a poor
country at the time in a union with Sweden – who was trying to come up with the
invention that Crookes had said, would be epoch-making for mankind’s progress.
Releasing nitrogen from the air requires great amounts of energy, and that
energy could be harnessed from Norway’s plentiful waterfalls. Electricity could be
produced more cheaply in Norway than almost anywhere else.
In the autumn of 1903, Sam Eyde and one of his partners, the Swedish
industrialist Knut Tillberg, campaigned to gain the interest of both German and
Swedish investors. Towards the end of the year, two Swedish investors, the halfbrothers Knut and Marcus Wallenberg, joined the project. This led to the
establishment of the company Elektrokemisk – Elkem, which is now a major
international supplier of metals and materials.
Notodden, in the south of Norway, was chosen as the site for taking the leap
from laboratory to factory. The area was ideal, as a hydroelectric power station had
already been built there in 1901. The nitrate company Notodden Salpeterfabriker AS
was established in Notodden in 1904, and from the autumn of that year, construction
was in full swing.
Outside a select circle, it was still unclear what the investors were expecting
from the Notodden plant. “Isn’t it an experimental unit?” asked a local newspaper.
The date is 24 September 1904, and the question is directed to the 26 year old
engineer Sigurd Kloumann, who is in charge of the construction job. “No,” he
answers. “It’s a factory.”
Production began in Notodden on 2 May, 1905, with around 100 workers
employed at the factory. This was envisaged as a first step towards a much larger
project, but these plans would soon change.
The foundation of Norsk Hydro should have taken place earlier, but these were
busy men who had to find time to meet. The papers were finally signed on 2
December, 1905 in Sam Eydes office in Christiania (now Oslo). The company was
named Norsk hydro-elektrisk Kvælstofaktieselskab later known as Norsk Hydro, or
just Hydro for short.
Figure 2-8. Norwegian engineer, Sam Eyde.
23
This was one of Eyde’s periods of glory. He was proud to relate that the
Notodden plants were now producing nitric acid, calcium nitrate and nitrite. “Calcium
nitrate, which has not previously been on the market, can be used in industry, and also
as Chilean saltpetre. Tests at various agricultural schools, in particular at Aas
Agricultural College by the head of the college, Mr. Sidelien, have shown that
calcium nitrate is as good as natural saltpetre and even somewhat better on sandy
soil.”
Waterfalls would be developed and plants would be built. In 1927 150
Electrolyser Units for Ammonia production have started at Rjukan, Norway. Installed
capacity was 30,000 Nm3/h, 150 MW.
Enormous sums would be invested. But there was no cause for concern. The
new industry would not market its products only in “our little country,” but all around
the world, as Eyde explained.
This date, 2 December, is still celebrated in the company every year as ”Hydro’s
birthday”.
In Denmark the first experiences with applied water electrolysis were obtained
by Poul la Cour [8], the leader of the folk high school Askov. He set erected the first
electricity producing windmill in Denmark in 1891. He soon became interested in
electrolysis in order to store the energy and use it for illumination at the school. After
some experiments with home made electrolyzers he got in contact with Professor
Pompeo Garuti in Italy who had invented an electrolyzer for use at a weapon factory,
and in 1894 he received the electrolyzers from Garuti. The produced hydrogen was
distributed in the buildings and used in special hydrogen lamps. As the hydrogen
flame is practically invisible, a refractory body was heated by the flame and visible
light was emitted from it. This way la Cour made what might have been the first
attempt to a local hydrogen society. He also experimented with hydrogen powered
combustion engines, but with less success.
24
3 Fundamentals of electrolysis
Many different types of electrolysis cells have been proposed and constructed.
The different electrolysis cells can be divided into groups based on the electrolyte.
Table 3-1 presents an overview of the different types of cells. All the cells presented
in Table 3-1 are capable of using H2O as reactant to produce H2. However, only the
solid oxide cell is capable of using CO2 to produce CO.
Table 3-1. Electrolysis cells and their specialties according to [9]
Type
Alkaline
Acid
Polymer
electrolyte
Solid
oxide
Charge carrier
OH-
H+
H+
O2-
Reactant
Water
Water
Water
Electrolyte
Sodium or
Potassium
hydroxide
Polymer
Ceramic
Electrodes
Nickel
Sulphuric or
Phosphoric
acid
Graphite
with
Pt,
polymer
Water,
CO2
Temperature
80 oC
150oC
Graphite
Nickel,
with Pt,
ceramics
polymer
80oC
850oC
The aim of this chapter is to give an overview of the thermodynamics and
general processes in an electrolysis cell.
In general, the electrolysis cell consists of two electrodes and an electrolyte. The
electrolyte may be a liquid (alkaline or acid) or a solid (polymer electrolyte or solid
oxide). It serves to conduct ions (the charge carrier) produced at one electrode to the
other. In order to avoid a short circuit inside the cell, the electrolyte has to be electron
insulating.
The overall electrolysis reaction (H2O → H2 + ½O2 or CO2 → CO + ½O2) is a
sum of two electrochemical reactions (also called half-cell reactions), which occur at
the electrodes. The electrode where the reduction of reactants or intermediates takes
place is called the cathode. The anode is the electrode where oxidation of reactants or
intermediates takes place. In Table 3-2 is shown the half-cell reactions for the
different types of electrolysis cells shown in Table 3-1.
In order to facilitate the electrode reactions, the electrodes have to be electron
conducting and catalytic active. Usually porous electrodes are used to maximize the
number of active sites in order to increase the activity and minimize material costs. In
most designs the porous structure is crucial to allow the reactants and products to
enter/exit the active sites in the electrode.
25
Table 3-2. Half-cell reactions for the different types of electrolysis cells.
Type
Alkaline
Acid
Polymer
electrolyte
Solid oxide
Charge
carrier
OH-
H+
H+
O2-
Cathode
reaction
Anode
reaction
H2O + 2e- → H2 + O22H2O + 2e- 2H+ + 2e- 2H+ + 2eor
→ H2 + 2OH- → H2
→ H2
CO2 + 2e- → CO + O2H2O → H2O
→
2OH- → H2O
½O
½O
+
+ O2- → ½O2 + 2e2
2
+ ½O2 + 2e2H+ + 2e- 2H+ + 2e-
3.1 Thermodynamics
3.1.1 Temperature
Both H2O and CO2 electrolysis become increasingly heat consuming with
temperature. Hence at elevated temperatures a significant part of the total energy
demand can be provided as heat according to Figure 3-1. This provides an opportunity
to utilize the Joule heat that is inevitably produced due to the passage of electrical
current through the cell. In this way, the overall electricity consumption and, thereby,
the H2 and/or CO production price can be reduced.
Figure 3-1: Thermodynamics of H2O and CO2 electrolysis at 0.1 MPa.
26
3.1.2 Thermo-neutral voltage
The thermo-neutral voltage is defined as:
ETn ≡
ΔH f
nF
=
ΔG f
nF
+
T ΔS f
nF
(4)
where ΔH f is the formation enthalpy change (see Figure 3-1), ΔG f is the Gibbs
free energy change, ΔS f is the formation entropy change, T is the temperature in
Kelvin, n is the number of electrons involved in the electrolysis reaction and F is
Faradays constant. Hence, if the cell voltage equals ETn all the produced Joule heat is
utilized. If the cell voltage is above ETn the cell produces surplus heat (waste heat).
If the cell voltage is below ETn, the produced Joule heat does not meet the heat
demand and the cell cools down if heat is not provided by other means.
For both H2O and CO2 electrolysis, ETn at 0.1 MPa, 25 °C is 1.48 V. At 950 °C,
it is 1.29 V and 1.46 V respectively. Hence, electrolysis of a H2O/CO2 mixture at 950
°C can be performed at thermo-neutral conditions at a cell voltage between 1.29 V
and 1.46 V depending on the H2O/CO2 electrolysis ratio.
3.1.3 Pressure
The equilibrium voltage, also called the reversible voltage or the electromotive force,
is determined by Gibbs free energy of water splitting and is thereby a function of both
pressure and temperature. The equilibrium voltage is the cell voltage at no current
load. It is equal to
ε=
ΔG f
(5)
nF
where ΔG f is the electric energy demand, see Figure 3-1.
ε
is also described by the Nernst equation
ε = ε 0 − RT ln
nF
PH2O
PH2 PO2
(6)
where ε 0 is the Nernst potential at standard pressure, R is the gas constant and T is
the temperature in Kelvin. PH2O is the H2O partial pressure, PH2 is the hydrogen partial
pressure and PO2 is the oxygen partial pressure. If the compositions of reactants and
products at the two electrodes differ, a cell voltage is established. Incidentally,
equals 0.93 V at 950 °C for both the H2O and CO2 electrolysis reactions.
It follows from equation (6) that ε increases with the overall pressure P:
ε0
27
Δε = −
RT
1
ln
,
nF
P
(7)
if we assume the overall pressure to be equal at both electrodes. At 850 °C an increase
in the overall pressure from 1 atm to 200 atm corresponds to an increase in ε by 0.13
V.
Although the cell voltage increases with pressure, pressurization may be
advantageous due to two main reasons: 1) Reduction of the internal cell resistance,
and 2) pressurization by means of heat (SOEC) or by an electrochemical overpotential
i.e. electrolysis with H+ or OH- conducting electrolytes, where the H2 is evolved at the
opposite electrode of the water supply.
In alkaline electrolysers pressure is known to reduce the internal resistance. In solid
oxide cells the impinging ratio of the gaseous reactant increases with pressure which
theoretically reduces the internal resistance. This has been experimentally verified
when the cell is operated as a fuel cell.[10,11] To the best of our knowledge, no
experiments with pressurized solid oxide electrolyser cells (SOECs) have been
published.
If the electrolysis reaction is performed high temperature, the pressure can be
achieved by heating (and thereby pressurizing) the inlet water. At 213 °C, the vapour
pressure is 20 atm. and at 287 °C it is 70 atm. Such "low temperature" heat may be
cheaper than electricity. Hence, the pressurization in the electrolyser may be cheaper
than pressurization by means of electric pumps as electrolysis cells are significantly
more efficient than pressurization pumps.
3.2 Efficiency
The efficiency, η , of the electrolysis process may be calculated as the higher
heating value (HHV) of one mole of the product divided by the energy consumption,
W, used to produce one mole of the product. W includes the electricity and heat used
for the electrolysis reaction plus energy losses of any kind, i.e.
η=
HHV
HHV
HHV
=
=
W
Electricity + Heat + Loss U ⋅ nF + Heat + Loss
(8)
where U is the cell voltage. HHV = ΔH fo where ΔH fo is the formation enthalpy
change at 0.1 MPa and 25 °C for one of the electrolysis reactions. If the electrolysis
cell is operated at or above ETn, all the heat for the electrolysis reaction is supplied by
Joule heat produced within the cell. Hence, equation (8) can be rewritten as
η=
ΔH fo
U ⋅ nF + Loss
=
ETno ⋅ nF
ETno
Loss
, L' =
=
U ⋅ nF + Loss U + L '
nF
(9)
28
ETno is the thermo neutral potential at 0.1 MPa and 25 °C. For all the cell types,
designs can be chosen, so that losses such as heat loss to the surroundings, electrical
leakage through the cell, gas leaks etc. are quite small. Hence, if the cell is operated at
ETno the cell can be operated with efficiency close to 100%.
A high efficiency is of course beneficial, however, an economically optimized
production is usually more important. In order to optimize the production economy a
high production rate is necessary. The higher the cell voltage is increased above ε the
higher is the current density and in turn the production rate. When the cell voltage
increases above ETno , surplus heat is produced and the efficiency decreases according
to (9). Today’s alkaline cells are typically operated at ~1.9V or higher in order to
optimize the production economy. In section 3.4 it is shown for SOECs that it is
possible to achieve economically optimized production costs and at the same time
keep the cell voltage at ETn or slightly above.
29
4 Types of electrolyzers
4.1 Alkaline electrolyzers
Alkaline electrolyte electrolyzers represent a very mature technology that is the
current standard for large-scale electrolysis. The anode and cathode materials in these
systems are typically made of nickel-plated steel and steel respectively. The
electrolyte in these systems is a liquid one based on a highly caustic KOH solution.
The ionic charge carrier is the hydroxyl ion, OH-, and a membrane porous to hydroxyl
ions, but not to H2 and O2 provides gas separation.
Key advantages of this technology include its maturity and its durability.
Key disadvantages are its use of a highly caustic electrolyte and its inability to
produce hydrogen at high pressures. This inability to produce high pressure hydrogen
for storage results in the added need for an external compressor, which adds cost and
complexity to the system.
4.1.1 The cell
In their basic design, water electrolyzers are quite simple Figure 1. They consist
essentially of anodes and cathodes isolated from one another by semi-permeable
membranes or separators, usually asbestos, all submerged in electrolyte, usually
KOH, held in some form of container. Direct current is passed through the cell and
water is decomposed to generate hydrogen on the cathodes and oxygen on the anodes.
The two gases are kept away from one another by the separators. The voltage drop
across the cell is a measure of its energy efficiency, i.e. the percentage of energy in
the electricity that is converted to hydrogen.
Figure 4-1. Alkaline water electrolysis cell.
30
When a current is passed through water, the molecules accept electrons from the
cathode, where their hydrogen's are reduced to H2 gas. The half-cell reaction is
4H2O(l) + 4e- → 2H2(g) + 4OH-(aq)
(10)
Other water molecules donate electrons to the anode, where oxygen gas is
produced:
2H2O(l) → O2(g) + 4H+(aq) + 4e-
(11)
The OH- ions and H+ ions produced by the electrolysis combine to produce
water again:
4OH-(aq) + 4H+(aq) → 4H2O(l)
(12)
and the net result is the breakdown of water to hydrogen gas and oxygen gas, with no
net change in the concentrations of H+ and OH-:
4H2O(l) + 4e- → 2H2(g) + 4OH-(aq)
2H2O(l) → O2(g) + H+ (aq) + 4e4OH-(aq) + 4H+(aq )→4H2O(l)
2H2O(l) → 2H2(g) + O2(g)
(13)
The hydrogen molecules accumulate on the surface of the
electrode until a bubble forms, breaks away, and rises to the surface of the electrolyte.
At the oxygen electrode, a similar process occurs in which hydroxyl ions are
discharged by giving up their electrons to the electrode and reacting to form water and
oxygen. The oxygen molecules accumulate into gas bubbles and rise to the surface.
Both of these electrode reactions require some intermediate catalytic reaction
with a metal surface. It is believed that the hydrogen ions discharge on the metal
surface to form an adsorbed layer of hydrogen atoms, which then recombine on the
surface to form hydrogen molecules. The ease with which the electrode reactions
occur is profoundly affected by both the physical and chemical natures of the surfaces
of the electrodes. A basic electrolyzer cell consists of the following components:
An Electrolyte: This is a water solution made conductive by mixing a salt or compound with water. Selection of the electrolyte is important because it must have the
following characteristics: It must exhibit high ionic conductivity; it must not be
chemically decomposed by voltage as large as that applied to the cell (so that only
water is decomposed); it must not be volatile enough to be removed with the evolved
gas; and, because hydrogen-ion concentrations are being rapidly perturbed at the
electrodes, the electrolyte should have a strong resistance to pH changes, i.e., it should
be a buffer solution.
For the most practical applications, these criteria can be met by the use of a
strong acid, such as sulphuric acid, or a strong alkali, such as potassium hydroxide
(KOH). Most salts are themselves decomposed under electrolysis at voltages likely to
be encountered in an electrolyzer cell. Acid electrolytes present severe corrosion
problems and are not usually selected for electrolyzers. Therefore, most commercial
31
electrolyzers operate with an alkaline electrolyte. Maximum conductivity occurs in
KOH solutions at about a 30% concentration, and this is the concentration usually
selected.
There is one notable exception to this use of alkaline electrolytes, the use of a
solid polymeric ion-exchange material that also has good ionic conductivity. Ionexchange resins having mobile negative ions (in other words, alkaline ion-exchange
resins) are notoriously sensitive to chemical degradation at elevated temperatures, and
this restricts the choice of ion-exchange electrolytes to acidic systems. The most
successful work with ion-exchange electrolytes has been carried out using a
polymerized fluorinated polystyrene sulphonic acid.
Electrodes That Have the Following Characteristics: They must be electronic conductors; they must have a suitable catalytic surface for the discharge of hydrogen or
hydroxyl ions; they must provide a large area interface between the catalyst and the
electrolyte; they must provide adequate sites for nucleation of gas bubbles; and they
must provide a reasonable means for the detachment of gas bubbles so that they may
separate themselves from the electrolyte at the operating voltage of the cell.
The form of the electrodes varies considerably from one cell design to another.
Large surface areas are obtained by the use of sintered structures, finned bodies,
screens, perforated plates, and flat plates with electrochemically roughened surfaces.
In the alkaline cells, nickel is the most commonly used catalytic surface. Rather than
making electrodes out of solid nickel, nickel-plated mild steel is often used. The
application of precious-metal catalysts, such as platinum, assists the electrode
processes considerably and allows them to proceed more rapidly than on nickel, but
the extra cost of the precious metal is not usually considered justified.
In the case of the polymeric acid electrolyte, electrodes must be made of more
chemically resistant materials than nickel or steel. Tantalum and gold have been used,
while the precious metals themselves, platinum, rhodium, iridium, etc., are usually
considered necessary as catalysts. When platinum is used, a large surface area can be
obtained by the use of so-called platinum black, a finely divided powder of platinum
metal particles.
A Separator: Required between the two electrodes, this serves the following purposes.
It prevents the electrodes from touching each other and shorting out, and it prevents
the hydrogen and oxygen gases from mixing together inside the cell. To provide this
function properly, the separator must consist of a porous diaphragm or matrix through
which the electrolyte solution can pass, affording an ionic conducting path from one
side of the cell to the other. These pores must remain full of liquid so that gas cannot
penetrate them. Additionally, the separator material must not be corroded by the
electrolyte in the presence of hydrogen or oxygen gas, and it must remain structurally
stable for the entire operating life of the cell so that the pores do not collapse.
To keep the ionic resistance of the cell as low as possible, the separator is
usually made in the form of a thin sheet, the thickness of which is determined by mechanical strength and gas crossover limitations. In the case of alkaline cells, asbestos
has commonly been used for the separator material. Woven asbestos cloth and matted
asbestos fibres are both used in commercial cells. Some experimental materials,
including potassium titanate, have been used in other alkaline cells. In the case of the
polymeric acid ion-exchange resin, this material acts as its own separator; and no
additional material is needed.
32
A Container: This serves to hold the electrolyte. In some cells, a nickel-plated steel
tank with a lid is used, while in others, solid metal sheets are interposed between the
electrodes, which are then stacked together with peripheral gaskets used to seal the
outer edges. This way, no separate container is required, and current is passed from
one electrode to the next through the metal separator plate.
4.1.2 Stacks and systems
In addition to the basic components of the electrolyzer cell itself, an electrolyzer
"system" requires further components. These include power-conditioning equipment
to convert ac power to the dc current required by the cell; electrical bus bar equipment
to distribute the dc power to the various electrodes in an assembly of electrolyzer
cells; gas-exit pipe work to duct the hydrogen and oxygen away from the cell;
separation systems to separate the gases from the electrolyte, which may be entrained
with the gas or deliberately circulated out of the cell with the gas; cooling systems to
remove waste heat from the cell itself; and drying systems to dry the hydrogen and
oxygen after they have been generated.
The electrolyte used in the conventional alkaline water electrolyzers has
traditionally been aqueous potassium hydroxide (KOH), mostly with solutions of 20–
30 wt % because of the optimal conductivity and remarkable corrosion resistance of
stainless steel in this concentration range [12].
The typical operating temperatures and pressures of these electrolyzers are 70–
100◦C and 1–30 bar, respectively.
Water electrolyzers traditionally have been grouped in two classifications-unipolar and bipolar [13,14].
The oldest form of industrial electrolysis of water uses the tank electrolyzer in
which a series of electrodes, anodes and cathodes alternately, are suspended vertically
and parallel to one another in a tank partially filled with electrolyte. Alternate
electrodes, usually cathodes, are surrounded by diaphragms that prevent the passage
of gas from one electrode compartment to another. The diaphragm is impermeable to
gas, but permeable to the cell's electrolyte. The whole assembly is hung from a series
of gas collectors.
A single tank-type cell usually contains a number of electrodes, and all
electrodes of the same polarity are connected in parallel, electrically, as pictured in
Figure 2.
This arrangement allows an individual tank to operate across a 1.9 to 2.5 volt dc
supply. In general, the cost of electrical conductors increases as the current load
increases, but the cost of ac-dc rectification equipment per units of output decreases as
the output voltage increases. This is one important consideration in the design of tanktype electrolyzers. There are two major advantages to tank-type electrolyzers:
(1) Relatively few parts are required to build a tank-type electrolyzer, and those
parts that are needed are relatively inexpensive. Because of this feature, tank-type
electrolyzers tend to optimize at a lower thermal efficiency than do more sophisticated
electrolyzer structures. Therefore, tank-type electrolyzers are usually selected when
electric-power costs are at their lowest;
(2) Individual cells may be isolated for repair or replacement simply by shortcircuiting the two adjacent cells with a bus bar. This feature allows maintenance to be
carried out with a minimum of downtime for the entire plant.
33
Figure 4-2. Schematic diagram of a unipolar (tank-type)
The major disadvantages of tank-type electrolyzers are:
(a) their inability to operate at high temperatures because of heat losses from
the large surface areas;
(b) their requirements for more floor space than other types of electrolyzers (a
disputed point); and
(c) the difficulty of designing the tanks to operate at high pressures.
The Canadian industrial program at the beginning of 1980s has selected the
unipolar design for development of advanced electrolyser technology13.
Electrolyzers of the bipolar design, on the other hand, may consist of a single
massive assembly of a relatively large number of electrodes, each of which is cathodic
on one side and anodic on the other (Figure 4). The assembly is held together by a
number of heavy longitudinal tie bolts, in a manner similar to that of the plate-andframe filter press. Each electrode is insulated from, and electrically in series with its
neighbour; and each pair of electrodes, with separating diaphragm, forms an
individual cell unit. The direction of current flow is from one end of the "cell pack" to
the other. A bipolar electrolyser may thus contain from twenty to several hundred
individual cells in series at 1.7-2.0V each, so that the corresponding applied voltage
ranges from 35 to 600V D.C., depending on required output capacity.
34
Figure 4-3. A bipolar stack.
Electrolyzers of the bipolar design, on the other hand, may consist of a single
massive assembly of a relatively large number of electrodes, each of which is cathodic
on one side and anodic on the other (Figure . Principle of a bipolar electrolyzer design
One advantage of the bipolar electrolyzer stacks is that they are more compact
than monopolar systems. The advantage of the compactness of the bipolar cell design
is that it gives shorter current paths in the electrical wires and electrodes. This reduces
the losses due to internal ohmic resistance of the electrolyte, and therefore increases
the electrolyzer efficiency. However, there are also some disadvantages with bipolar
cells.
One example is the parasitic currents that can cause corrosion problems.
Furthermore, the compactness and high pressures of the bipolar electrolyzers require
relatively sophisticated and complex system designs, and consequently increases the
manufacturing costs. The relatively simple and sturdy monopolar electrolyzers
systems are in comparison less costly to manufacture. Nevertheless, most commercial
alkaline electrolyzers manufactured today are bipolar.
As an alternative to tank-type electrolyzers, more recent electrolyzer designs use
stacks so that the positive electrode of one cell is directly connected to the negative
electrode of the next. An assembly of cells has superficial resemblance to a filter press
because the electrolyte is manifolded to flow through each cell in parallel while
hydrogen and oxygen exit lines are similarly manifolded through the stack.
Figure 5 is a schematic of a filter-press cell construction. This type of cell is
sometimes called a bipolar cell (in contrast to the monopolar assembly in the tanktype cell) because each electrode is used with one face as the positive electrode of one
cell and the opposite face as the negative electrode of the next cell.
35
Figure 4-4. Filter-press (bipolar) cell construction.
In practice, filter-press-type cells are usually constructed with separate electrodes in
each cell that are electrically connected through a solid metal separator plate that
serves to keep the hydrogen cavity of one cell separate from the oxygen cavity of the
next. Because the cells of the filter-press-type electrolyzer can be relatively thin, a
large gas output can be achieved from a relatively small piece of equipment. It is
usually necessary to cool the cells by circulating the electrolyte through them, and the
electrolyte exiting from the cell carries with it the gas produced. In many designs,
separation of the gas from the electrolyte is accomplished in a separating drum
mounted on top of the electrolyzer. The electrolyte, free of gas, is re-circulated
through the cells.
The major advantages of filter-press-type electrolyzers are that:
(a) they take up less floor space than the tank-type design;
(b) they are more amenable to operation at high pressures;
(c) they are more amenable to operation at high temperatures.
The major disadvantages are that:
(a) they require a much closer tolerance in construction because of sealing
problems, and;
(b) they are more difficult to maintain because if one cell fails, the entire
battery has to be dismantled and production of hydrogen is lost.
Filter-press electrolyzers usually present higher capital costs per unit area than
tank- type cells, and, to compensate for this, they are operated at higher current
densities.
The bipolar design has been universally accepted as offering the most potential
for incorporation of advanced technology [15- 19]. This is in part due to the intangible
attractions of the highly-engineered bipolar cells, and to the recent development of
new bipolar concepts for application in cost-insensitive aerospace applications. Also,
cost and performance projections for bipolar electrolyser equipment have been
erroneously compared with data for early unipolar plants, which are not representative
of the potential of the unipolar approach.
There is, in fact, a clear distinction between unipolar electrolyser technology
[20- 23] and the older designs on which most published comparisons with bipolar
equipment have been based.
In new advanced alkaline electrolyzers the operational cell voltage has been
reduced and the current density increased compared to the more conventional
electrolyzers. Reducing the cell voltage reduces the unit cost of electrical power and
thereby the operation costs, while increasing the current density reduces the
1 1 1
2 2
36
investment costs [12]. However, there is a conflict of interest here because the ohmic
resistance in the electrolyte increases with increasing current due to increasing gas
bubbling. Increased current densities also lead to increased overpotentials at the
anodes and cathodes.
Three basic improvements can be implemented in the design of advanced
alkaline electrolyzers:
(1) new cell configurations to reduce the surface-specific cell resistance despite
increased current densities (e.g., zero-gap cells and low-resistance diaphragms),
(2) higher process temperatures (up to 160 oC) to reduce the electric cell
resistance in order to increase the electric conductivity of the electrolyte, and
(3) new electrocatalysts to reduce anodic and cathodic overpotentials (e.g.,
mixed-metal coating containing cobalt oxide at anode and Raney-nickel coatings at
cathode).
In the zero-gap cell design the electrode materials are pressed on either side of
the diaphragm so that the hydrogen and oxygen gases are forced to leave the
electrodes at the rear. Most manufacturers have adopted this design14.
Cell Container
The material qualities required for the cell container are: high chemical resistance,
high mechanical strength, good insulation properties, good machinability and a cost as
low as possible. In the case of water electrolysis, steel (mild steel) is more frequently
used as the container material for the cells. Parts in danger of corrosion are protected
by nickel plating or by other cheaper insulation materials such as rubber, compressed
asbestos, ebonite, alumina, cement or Teflon.
On the other hand, synthetics such as polyethylene, a number of types of nylon,
and epoxy resins have so far proved to have an inadequate working life under the
conditions of electrolysis [24].
Electrolyzer-System Designs
A total electrolyzer system consists of all the equipment necessary for the process,
from the input of electrical power to the output of hydrogen and oxygen gas at the
appropriate purity and pressure levels. In addition to the electrolyzer cell module
itself, which has already been described, three major subsidiary systems can be used
in various forms.
Power Supply: For relatively large-scale electrolyzer systems, power is usually
supplied from a three-phase, high-voltage line. To convert this into the relatively lowvoltage dc power needed for the electrolyzer cell, a combination transformer rectifier
unit is usually used. There is a trade-off to be made in the design of the transformerrectifier system, which can provide dc at relatively high or relatively low voltages. By
connecting the cells in series, high-voltage dc systems can be used, and this can have
some cost advantages in the requirements for transformers and rectifiers.
For reasonably large systems, dc voltages of 70 to 100 volts are usually used.
Clearly, this is not possible with very small units because a large number of very
small cells would be needed.
The cost of a transformer-rectifier system is considerable and can represent as
much as one-third to one-half of the cost of the entire system. If electric power is
being generated onsite, some consideration should be given to the direct generation of
dc power and to the use of this for electrolysis. There seem to be no examples of this
in other electrochemical installations, for example, in chloride caustic plants or
37
aluminium-smelting installations that use on site power. However, recent
developments in the technology of acyclic or dc generators may make the direct
reduction of dc power more promising. Modern, acyclic dc generators operate only at
low voltages and this implies the use of very large currents, very large bus bars to
distribute the power to the electrolyzer cells, and very complicated switch gear for
handling high-current, low-voltage dc.
On the other hand, dc generators apparently can be produced for about the same
cost as ac generators; and the use of the dc system could considerably reduce capital
costs that would otherwise be required in the provision of transformer rectifier units.
At present, not enough information is available to draw any conclusions about the
relative merits and disadvantages of the ac versus dc supply systems.
Cooling Systems: Because electrolyzer cells are not, in fact, 100% efficient, a
considerable amount of waste heat is generated in the electrolyzers and must be
removed from the cells. There are several ways of doing this: (a) by circulating
electrolyte, (b) by circulating hydrogen, (c) by circulating water through the cell, and
(d) by circulating water through a heat exchanger in contact with the cell.
Circulation of electrolyte requires a pump capable of handling a corrosive liquid
at relatively high temperatures and possibly at a high pressure. If electrolyte is
circulated through a common manifold through a large number of cells connected in
series, then a high voltage is applied to it from one end of the manifold to the other.
This induces a short circuit through the electrolyte, thus utilizing only the electrodes
at either end of the cell stack. There is a trade-off between the reduction of this shortcircuit current or "shunt current," which results in low current efficiency of the entire
cell stack, and the deliberate introduction of high-resistance paths in the electrolyte
circulation loop, which result in a requirement for high circulating pumping power.
In some types of cells, notably the tank-type cells in which the electrolyte in
each cell is kept entirely separate from that in all others, these shunt currents are not
possible. The circulation of electrolyte in these cells is usually provided by the gas-lift
effect of the gases being evolved at the electrodes. Thus, very little parasitic energy is
required, and no electrolyte circulating pump is needed. However, the circulation rates
achieved by this means are not usually sufficient to remove the generated heat from
the cell, but simply serve to stir up the electrolyte to reduce concentration gradients
resulting from the removal and replacement of water.
Hydrogen itself can be used as a heat transfer material by circulating it repeatedly through the cell. Again a circulating pump is required that can handle hydrogen,
sometimes in the presence of traces of electrolyte. Hydrogen is withdrawn from the
circulating loop at the rate at which it is produced at the electrode, and the circulating
loop contains the heat exchanger by which the waste heat of the cell is removed.
In the SPE-type cell, it is possible to circulate water through the cell, in contact
with the electrolyte, without leaching out the electrolyte itself. This approach is not
possible in a cell that uses an aqueous electrolyte solution; and, in this case, a separate
water compartment must be used. This is easier to achieve in a tank-type cell than in a
filter-press type, although water-cooled plates can be built into stack-type cells. In
some tank cells, a water chest, to act as a heat-removing mechanism, is incorporated
into the design of the tank itself.
One of the problems of operating electrolyzer cells at very high pressures is that
the auxiliary equipment, including the cooling system, would also have to be operated
at high pressures; and thus the cost of even electrolyte and feed-water pumps, which
in an atmospheric system would be insignificant, can become considerable.
Gas-Removal Systems: Once gas has been generated at the electrodes, it must be
38
removed from the electrolyzer cell and conditioned to the temperatures, pressures, and
purity levels required by the customer. There are two ways of removing the gas from
the cell. One is to allow it to be entrained in the flowing stream of electrolyte,
bringing both out from the cell together, and passing the stream through an external
separator. This usually makes the design of the electrolyzer cell itself simpler, but
requires extra equipment for the separation of electrolyte from the gas. Clearly, two
separator systems would be required, one for hydrogen, and one for oxygen.
The second method is to allow the gas to separate itself from the electrolyte
within the cell and then remove it as a gas stream only. In this case, it is likely to
carryover a spray of electrolyte, and a spray trap of some sort is needed. Once
hydrogen and oxygen have been removed from the cell, they must be dried because
they are produced from the cell saturated with water vapour. After drying, they must
be compressed if the cell is not operating at the required delivery pressure. This need
for an external compressor increases the parasitic load or energy requirement of the
overall cell system.
The removal of small traces of oxygen from the hydrogen stream can be accomplished by use of a so-called "deoxo" catalyst. This is usually a high-surface area
palladium catalyst, supported on asbestos, which has the effect of causing the traces
of oxygen to combine with hydrogen to form water. Because oxygen and water
vapour are the only major impurities likely to be found in electrolytic hydrogen,
drying and oxygen removal are the only purification steps necessary for obtaining
very high purity hydrogen.
4.1.3 State of art
Separator materials for use
in alkaline water electrolyzers
Diaphragm
The purpose of the diaphragm in an electrolysis cell which produces gases at either
electrode is threefold [25]:
(1) The diaphragm has to prevent unhindered intermixing of catholyte and
anolyte. The gas evolution at both electrodes forms a two-phase mixture of electrolyte
with more or less dispersed bubbles so that intermixing of anolyte and catholyte
always means intermixing of the two gases which should be prevented strictly in
order to obtain high gas purities and current efficiencies, respectively.
(2) The diaphragm must form an efficient diffusion barrier for the gas molecules
in order to prevent contamination of the evolved gases by molecular diffusion of the
gas which is generated at the respective counter electrode.
(3) The diaphragm may be used further to prevent, very efficiently, the
formation of a gas bubble curtain at the front side of the electrodes just by pressing
the electrodes onto the, more or less, elastic diaphragm.
Most important for all three purposes is that clogging of the diaphragm pores by
gas bubbles, which may either intrude into the pore mouths or which may precipitate
out within the pores from gas-supersaturated electrolyte solutions, must be excluded
completely unless the electrical resistance of the diaphragm increases in an
uncontrolled manner. Bubble formation in small cavities, pores, etc. of radius r may
only be observed if a certain degree of supersaturation is established:
39
P –Psat. ≥ 2σ
r
(14)
Owing to gas evolution at the back side of the electrodes, i.e. in greater distance
from the diaphragm, the supersaturation of the gas in the electrolyte which contacts or
diffuses or drifts into the diaphragm pores will be largely reduced. Under a working
pressure of 30—60 bars, supersaturation pressures of hydrogen and oxygen of no
more than a few bars will exist at the diaphragm surface*. According to equation (14)
for surface tensions of the electrolyte of approx. 200 dyn cm-1, pore diameters of some
micrometers will prohibit gas clogging of the diaphragm reliably.
The diaphragm must additionally offer a sufficiently high hydrodynamic
resistance to retard intermixing of oxygen saturated anolyte with hydrogen saturated
catholyte due to occasional pressure differences between the cathodic and anodic
compartments. By knowledge of the solubility of oxygen and hydrogen in caustic
potash the maximally permissible fluid dynamic permeability of a diaphragm is
estimated [26] to be approx.
k
max
≈ 5 cm3 centipoises (cm2 bar s)-1,
hydro
(15)
in 50 wt % caustic potash at 60 bars for i = 1 A cm-2.
According to this estimation a relatively high surface specific hydrodynamic
resistance of the diaphragm is necessary. Nevertheless, the diaphragm must offer only
a low electrical surface specific resistance if immersed in the electrolyte. The surface
specific electrical resistance must not exceed 0.2 Ω cm2 and should rather be around
150-100 m Ω cm2 in order to avoid too high ohmic potential drops within the
diaphragm at current densities around 1 A cm-2. Because of the thermodynamic
instability of asbestos (chrysolite) in caustic potash [27] it seems necessary to develop
a completely new diaphragm material which is based on a porous refractory structure
and which, additionally, would allow to modify the diaphragm properties at will.
A good diaphragm should present a low resistance to the flow of current and as
far as possible be chemically resistant to the electrolyte.
Traditionally asbestos cardboards, papers and woven cloths have been used in
alkaline electrolyzers. Increase of the electrolyser temperature above 100°C makes
this material unsuitable.
Inorganic materials
Asbestos
Resistance of asbestos to high-temperature electrolyte. The chemical and physical
stability of asbestos in alkaline media at high temperatures has been studied by
several groups of researchers [28- 32] . J. W. Vogt [28] has examined the chemical
degradation of fuel-cell-grade chrysotile asbestos in KOH solutions of 30, 40, 50 and
60 wt. % for 100- and 1000-h periods at temperatures of 50, 100, 150 and 200 °C.
Sample dissolution was found to increase with increasing time, temperature and KOH
concentration. The weight losses attained an upper limit of 40 % except for a few
samples. Leached samples retained some of the fibrous structure of asbestos. Large
amounts of granular material, brucite (Mg(OH)2), were scattered in the leached
2 3 3
40
samples and filled the remaining fibre clusters. Vogt's report concluded that chrysotile
asbestos is not a satisfactory material for extended service at temperatures as high as
100 °C, but that the insoluble residue, if it retains a satisfactory structure, might be
superior to asbestos. Results of alkaline leaching of other types of asbestos, tremolite,
amosite, anthophylite and crocidolite, were similar to those obtained with chrysotile.
Leaching treatments of chrysotile and of the other asbestoses with HCI or with
sequestering agents gave poor results, leaving behind nearly-pure silica fibres which
dissolved readily in alkali.
Asbestos stabilization. Crandall and Harada [29] have attempted to improve the
resistance of chrysotile asbestos to potassium hydroxide at temperatures greater than
100°C. They have identified the asbestos-KOH reaction and studied the effects of
environmental parameters on its rate. A highly-converted asbestos matrix was
developed. Coating and stabilization studies were also carried out.
The reaction between asbestos and KOH at temperatures up to 200 °C was
established to be
Mg3Si2O5(OH)4 + KOH(aq) = 3Mg(OH)2 + Soluble silicates (aq).
Debye-Scherrer powder patterns of the residues showed both chrysotile and
brucite phases to be present. The deterioration of asbestos was demonstrated to be
caused by the leaching of silicon from the asbestos, leaving behind insoluble brucite
and soluble potassium silicates. The leached samples retained a fibrous morphology
as if the Si-O layers were stripped off, leaving the Mg-OH structure intact. This
"converted" asbestos exhibited improved chemical stability suggesting the use of such
materials as matrices.
The conversion of asbestos fibres to flexible and chemically-resistant bruciteasbestos fibres was best achieved by a repetitive leaching-washing procedure.
Composition as well as morphology of the converted material was dependent on the
leach-wash cycle conditions; a typical composition of the converted material was 80
% Mg(OH)2 and 20 % asbestos. A product with satisfactory properties was obtained
by leaching asbestos in 40 % KOH at 150 °C for five 20-hr cycles. Mats formed with
the converted asbestos were tested by NASA Lewis Research Center and Pratt and
Whitney. The latter found that these mats had unacceptably-low bubble-pressures and
lacked stability upon exposure to KOH at 120 °C. Similarly, attempts to fabricate
brucite matrices were unsuccessful [33].
Crandall and Harada [29] hypothesized that asbestos could be stabilized by
addition of suitable amounts of soluble potassium silicate. Results of corrosion studies
in 40 % KOH at 150 and 200 °C indicated that the corrosion reaction was essentially
eliminated with a potassium silicate to potassium hydroxide weight ratio of 1:10, and
of 1.5:10 for 60 % KOH solutions. X-ray analyses and S.E.M. examinations of the
stabilized samples exposed to 40 % KOH at 200 °C for 20 h have shown no change in
the chemical and physical structure of the asbestos. The above stabilization process is
the subject of U.S. Patent No. 3,891,461 [34].
Harada and Crandall also tried to improve the caustic stability of asbestos by
application of a reaction barrier. Attempts to apply TiO2 and ZrO2 by chemical vapour
deposition from tetraisopropyl titanate and tetraisopropyl zirconate precursors were
unsuccessful. Weight losses for the ZrO2- coated asbestos were about the same as for
uncoated asbestos, while those for the TiO2-coated asbestos were somewhat higher.
This approach was abandoned in favour of the stabilization and conversion processes.
Potassium titanate
41
Fibrous and acicular alkali-resistant potassium titanates have the general formula
K2O(TiO2)n. Potassium tetratitanate (n = 4), hexatitanate (n = 6), and octatitanate (n =
8) have been characterized by X-ray diffraction. The tetra- and octatitanates are much
more fibrous than hexatitanate. Chemical stability is greatest for the hexatitanates and
least for the tetratitanates. The use of these materials as separators in alkaline
electrochemical systems has been envisaged for some time by several groups [28,3538].
Vogt [28] has shown P.K.T. (pigmentary potassium titanate)
(K2O(H2O)x(TiO2)10-13) to be stable at temperatures as high as 150 °C, in 30 to 50 %
KOH electrolyte. However, P.K.T. and other potassium titanates such as Fybex or
Tipursul, originally fabricated by DuPont, cannot be formed into matrices without a
binder. Separators fabricated by Vogt with Teflon extrusion powder as a binder
showed generally good stability at temperatures as high as 150 °C. These diaphragms
were described as being "matrices of inorganic powder incorporated in a network of
very fine Teflon fibres". The latter were wettable and had fair gas-sealing properties.
Vine and Narsavage [33,36] have also prepared potassium titanate matrices,
using Fybex from DuPont. The Fybex material is (K2O)x(TiO2)z, with z/x equal to
about 8. Fibre diameters were in the range 0.10-0.15 /µm, and typical fibre lengths
were 4/um with a range of 2-22/µm. No evidence of deterioration was detected for a
mixture of 4 % Teflon-3170 and 96 % Fybex, after 3000 h at 120 °C in 42 % KOH.
100 % Fybex matrices were prepared but were fragile. Low-bubble pressures were
obtained for Fybex compositions bonded with Teflon-3170, screen-printed onto
different types of electrodes with several different variations in printing techniques.
This approach was abandoned. Another separator fabrication method was preferred
which consisted of filtering a 96 % Fybex plus 4 % TFE-3170 water-based
composition directly onto fuel-cell electrodes, followed by drying and partial curing
of the matrix binder. Bubble pressures of 0.21-0.23 MPa were obtained for 0.75-mm
matrix coatings having 90 % porosity. The latter matrices resisted a test of 2300 hours
at 120 °C with no deterioration. No further tests were done on these matrices, as
DuPont had stopped production of Fybex. The authors tested beta silicon nitride
(Si3N4) as an alternative without success.
Post and co-workers from NASA Lewis Research Center [37] have examined
representative forms of potassium titanate for durability in hot caustic, including
fibrous tetra- and octatitanates, fibrous variations of tetratitanates and acicular and
non-acicular hexatitanates. The purpose of this work was to relate the parameters of
the material (i.e. size and shape of particles, chemical and crystallographic
characteristics) to its performance in matrices. Ways of improving performance and
test methods for screening were explored. However, this work did not include the
fabrication of matrices. It was found that P.K.T. is somewhat susceptible to chemical
degradation, although it is far superior to asbestos. P.K.T. was considered to be
unsuitable for fabrication of matrices due to existence of the fibres in many forms and
to the presence of non-fibrous and colloidal matter. Octatitanates were judged to be
more suitable. The author also suggested that alkali induced fibrillation of potassium
titanate crystal clusters might be used to obtain fibres with lengths in the millimetre
range. The necessity of careful dispersion of the titanate fibres in separator fabrication
processes was underlined due to the tendency of the fibres to form clusters. Finally, it
was emphasized that a fractionation procedure is necessary for evaluation of the
potassium titanates, due to the wide range of shapes and sizes which characterizes this
material.
3 3
42
The electrochemistry division of Compagnie Generale d'Electricite, France has
fabricated experimental separators of potassium titanate fibres bonded with P.T.F.E.38.
A suspension of stabilized Teflon was mixed with a lubricated paste of potassium
titanate fibres. The above mixture was calendared to a thickness of 300µm and dried
in air at 160 °C. 30-40 % Teflon is optimum for obtaining good mechanical properties
and low resistivity. Best results were obtained with fibres of 0.5 µm diameter and 5
µm length. In order to improve further the mechanical properties of the membrane, it
was pressed between two nickel expanded-metal sheets. The resistance of a 300 µm
membrane, measured in 34 % KOH at 25 °C was 0.2 to 0.3 Ωcm2 [38a,c]. This is
equivalent to 0.1 to 0.2 Ωcm2 at 120 °C. Values as low as 0.09-0.10 Ωcm2 had been
reported earlier by Appleby and Crepy [38b].
In conclusion, potassium titanates show excellent stability in hot caustic
environment. Post [37] suggested that the octatitanates offer the best combination of
properties for use as a matrix in alkaline media. Matrices fabricated from potassium
titanate alone are too fragile. It is necessary to use a binder such as P.T.F.E. Such
matrices have shown excellent ionic conductivity.
Polyantimonic acid
A membrane, based on an inorganic cationic ion-exchange material,
polyantimonic acid (PAM), has been developed by a Belgian laboratory for use in
alkaline water electrolyzers [39- 43]. The Belgian group had tested other ionexchange materials before selecting polyantimonic acid; the latter was preferred over
zirconium phosphate, zirconium oxide, tin oxide, bismuth oxide and lead sulphate
[43].
Polyantimonic acid was first prepared by Baetsle [39]. The unit cell has the
following rather complicated empirical formula [H3Sb3O5(OH)8]3 [H5Sb5O6(OH)18].
Fourteen atoms of hydrogen are exchangeable in theory, corresponding to an
ion-exchange capacity of 5.05 meq g-1. An ion-exchange capacity of 2.3 meq g-1 was
found for K+ ions at pH = 7, when the material was incorporated in a binder. The ionexchange capacity was only slightly higher for free PAM. PAM has been found to be
extremely stable in concentrated alkaline solutions at temperatures to 150 °C.
Membrane specimens were initially prepared by a low-temperature sheet-rolling
technique using Teflon 6 as the binder. The weight ration of PAM to P.T.F.E. was
80:20. Scale-up of the cold-rolling technique for fabrication of sheet sizes of 30cm x
30cm was not successful. A wet-agglomeration technique and other techniques using
Teflon fibres, or combinations of Teflon fibres with Teflon powder, were also
unsuccessful. Finally, a film casting technique was determined to be appropriate.
In the film casting technique, the binder is dissolved in a suitable solvent, and
the inorganic material is dispersed in this solution. A film is cast on a glass plate, and
the solvent evaporated. Polyvinylidene fluoride (P.V.D.F.) was tested as a binder, but
could not resist chemical degradation in the alkaline media at high temperatures.
Deterioration of the PAM-P.V.D.F. membrane occurred at temperatures above 60 °C.
Polyarylethersulfone-PAM (1:2 weight ratio) films were prepared, and were
expected to have much better chemical resistance. The membrane resistance of the 0.1
mm thick films dropped from 0.7 Ωcm2 at 25 °C to 0.2 Ωcm2 at 95 °C. The
fabrication process was reasonably reproducible. Tests conducted in 50 % KOH at
120 °C for 200h have indicated no dimensional changes or chemical deterioration for
the polysulfone-PAM membrane. Gas-tightness measurements showed a maximum of
1 % hydrogen in the oxygen stream when the membrane was tested in an electrolyser;
some improvements are needed in this area. Nevertheless, the polysulfone4 4 4
43
polyantimonic acid membranes must be regarded as serious candidates for waterelectrolyser use. Further long-term testing in electrolyzers is necessary to confirm the
suitability of the material for commercial use.
Other inorganic separators
Refractory-type materials. Various inorganic materials, alone or combined with
asbestos or an organic binder, were considered as possible separator materials for
alkaline electrochemical cells by NASA Lewis Research Center in 1968 [28]. Their
study included the following materials: ceria (CeO) powder, zirconia E fibre
(neodymia stabilized), TX fibre-nemolite (hydrated magnesia), and pigmentary
potassium titanate (P.K.T) from DuPont, boron nitride, beryllium oxide, zirconium
silicate, titanium dioxide, and silicon carbide. The following behaved poorly in tests
conducted at 100 and 150 °C: boron nitride, beryllium oxide, zirconium silicate, and
silicon carbide. Ceria and zirconia E fibre showed excellent stability under the test
conditions and were chosen as candidate materials. P.K.T. and titania were also
retained. TX fibre, although it had good stability, was rejected because of its
excessive iron content.
The above-selected materials were fabricated into separators by NASA Lewis
Research Center, alone or in combination with asbestos. Results with both the
composite separators and the inorganic materials alone were disappointing. The
fabricated separators were fragile, had poor strength and lacked cohesion when wet.
Pressed mats, formed by admixture of Teflon fibres or Teflon emulsion and the
preferred inorganic materials, also failed to meet the requirements for good
electrochemical separators. The resulting matrices were brittle when dry and semifluid when wet.
Another preparation method, based on the tendency of Teflon 6 extrusion
powder to convert to fibres under mild shear forces, was tested. The matrices were
prepared by blending the Teflon powder with the inorganic materials and sufficient
mineral spirits to form mobile slurry. The slurry was filtered on a Buchner funnel,
rolled with a rolling pin until the fibre development was judged satisfactory, and then
moulded to the desired shape. The optimum Teflon concentration was about 5 % by
weight with P.K.T. and 1 % for ceria, zirconia and magnesia.
The shearing of the Teflon 6 powder produced a network of very fine fibres,
distributed through the bulk of the matrices. The resulting porous mats were flexible,
strong and wettable, but were inferior to chrysotile in gas sealing capability and
electrolytic resistance. These matrices were nevertheless considered promising. No
data are known to have been published on the results of tests performed on full-size
matrices.
Hausmann [44] has reported preparation of ceramic separator materials which
incorporate plastics as binders. No performance results have been presented.
Sintered nickel. Some authors have recently suggested the use of porous
metallic diaphragms such as sintered nickel as separators in alkaline electrolyzers [4548]. These materials are highly resistant to corrosion, and give good gas purities and
ionic-conductivity. Sintered nickel plates, 0.6 mm thick, were tested in 30 % KOH at
pressures of 50 bars, temperatures higher than 150 °C, and current densities greater
than 20 kA m-2 [46]. The major problem with this type of diaphragm is high electronic
conductivity, necessitating complete insulation from the electrodes and the
electrolyser structures. Failure to achieve this could lead to the diaphragm acting as an
electrode (anode or cathode), resulting in the production of an explosive gas mixture.
P.T.F.E. spacers and seals were apparently used with some success. Divisek et al.
4 4
44
have reduced the electronic conductivity of the nickel structure by converting the
nickel surface to an oxide using thermal oxidation in air at high temperatures (>1000
°C)48. The electronic resistance of one diaphragm was increased from approximately
10-4 Ωcm to 107 Ωcm, while its resistance to the passage of ions in the electrolyte
increased slightly. Such a separator was tested successfully at 110 °C for 3000 h in an
electrolyser [48].
Metallic diaphragms have several desirable properties, and it seems to be
possible to eliminate the problems associated with their electronic conductivity.
However, such diaphragms may prove too costly for wide-spread use. Perroud cited a
price of $1000 m-2 for nickel sinter, with a possible price-reduction factor of 5 for
large quantities46. Prices of other porous metallic diaphragms should be critically
examined.
Oxide-coated nickel materials. Fischer et al. have suggested coating of nickel
"nets" with aqueous slurry of corrosion-resistant ceramic materials, followed by
drying and sintering [47]. For example, aqueous slurries of BaTiO3 and BaTiO3 (46%)
+ ZrO2 (46%) + K2Ti6O13 (4%) + Na2TiO3 (4%) were tested with other mixtures. The
ceramic coating so obtained insulates the nickel "net" on both sides. This thin coating
would be flexible enough to be bent without cracking around a radius of 3 cm. Ohmic
resistances as low as 0.027 Ωcm2 to 0.054 Ωcm2 were obtained for such diaphragms
in 30 % KOH at 25 °C. The diaphragms were tested in electrolyzers at 252 °C and 35
bar pressure, in 50 % KOH. However, no long-term test results were available. These
separators might be less expensive than nickel sinters.
One group has proposed the fabrication of separators by plasma spraying of
inorganic materials, such as ZrO2, MgO and TiO2, on a supporting metal grid or mesh
[49]. Ceramic oxides have outstanding chemical resistance in alkaline media; they can
not be used unfortunately in the form of thin sintered sheets as these are too brittle and
sensitive to thermal shocks. This led to the development of plasma-sprayed inorganic
(PSI) membranes, which are fabricated by spraying a molten refractory oxide onto a
suitable substrate, with a plasma spray gun. The sprayed material can be coated on the
electrodes themselves or on another support. It is possible to control the parameters of
the resulting coatings, including thickness, porosity and pore-size distribution. Oxides
such as MgO, TiO2 and ZrO2 could be used, according to the authors; MgO was
reported however, to convert in hot caustic to Mg(OH)2 which precipitates out of the
solution [47]. The supports can be fine- or coarse-mesh gauge, sintered porous sheets,
or expanded-metal.
The above processes could produce membranes having excellent properties for
use in high-temperature alkaline electrolyzers. Some oxide-coated membranes have
been shown to have lower resistivities than polyantimonic acid and potassium
titanate47. It has yet to be demonstrated, however, that fabrication will be possible at a
reasonable cost. Also, satisfactory performance must be demonstrated in long-term
electrolysis tests.
Organic materials
Organic polymers could provide attractive alternatives for the replacement of the
currently-used asbestos, as they can be spun into fibres, and these in turn can be
prepared in woven cloths, felts or other non-woven fabrics. Organic polymers could
also be prepared as microporous films. Although modern polymers lend themselves to
a variety of processes suitable for the preparation of different forms of separator
materials only a very few can survive the environmental conditions existing in an
alkaline water electrolyser at 150 °C or above.
45
Some of the polymers suitable for use in water electrolyzers have been tested by
Teledyne Energy Systems for fabrication of both structural parts and separators [50,52]. These included polyphenylene sulphide and the polysulphones.
The polysulphones
The polysulphones are a family of polymers which have in common sulphone groups
(---S---) which provide thermal stability and high temperature rigidity, and ether
groups for toughness [53]. Three main types of polysulphones are available: Udel
[54a] , a polyarylethersulphone; Radel [54b], a polyphenylsulphone; and Victrex
[54c], a polyethersulphone. Astrel 360, a polyarylsulphone which was introduced by
3M Company, has been phased out of production by the Carborundum Company
because of high cost and poor processability. Finally, Mindel A-650, modified
polysulfone, has been introduced by Union Carbide. The polysulphones possess
excellent thermal and oxidative stability. All bonds in the polysulfone structure should
be hydrolytically stable, making them stable in alkaline or acid environments even at
high temperatures.
Reinforced and non-reinforced polysulphones were tested by Teledyne Energy
Systems for use in structural parts [50]. The non-reinforced polysulfone behaved
satisfactorily in 25 % KOH at 82 °C. After 50 weeks, a small weight and dimension
change attributed to water absorption was observed. Teledyne contracted Fabric
Research Laboratories of Dedham, Massachusetts to prepare fibres of the other
polysulphones, and tested porous matrices prepared from the latter by exposure to
KOH/O2 and KOH/H2 environments at 150 °C [51] . Solution-spun polyarylsulphone
fibres of 12 µm diameter dissolved completely within 500 h, while 16 µm polysulfone
(Udel) fibres lost some strength.
Results of the above tests led Teledyne to recommend maximum temperatures
of 125 °C for polysulfone (Udel), and of only 100 °C for polyethersulphone (Victrex)
and polyarylsulphone (Astrel 360). All tested matrices had inadequate gas retention
due to too-large pore diameters and poor wettability of the fibres. Approaches
suggested for resolution of this problem were:
(i) the use of finer fibres in the preparation of the matrices;
(ii) an increase in the polymer critical tension through modification of the
surface; and (iii) addition of wetting agents to the electrolyte [52].
British Patent No. 1,435,420 [55] describes a fabrication method for
polysulfone fibres with diameters between 0.01 and 21 µm. Some matrices were
fabricated, according to the patent, with 67 % of the fibres having diameters between
0.25 and 1.0 µm. Another way of improving the wettability of the fibres would be to
modify the polymer, for example, by addition of anionic groups such as (SO3-) [56] or
-COOH.
Ultra filtration membranes and anisotropic membranes have been prepared by
different companies. One of these, Osmonics Inc., has ultra-filtration membranes
available with 25-50 µm thickness, and with maximum pore sizes of 50-100 Å. This
is too small a pore size to provide a good wicking of the electrolyte in the separator;
the result would be excessively-low ionic conductivity.
In summary, the polysulphones are polymers having excellent thermal,
oxidative and hydrolytic stability. However, when prepared as fibres, their maximum
service temperatures in water electrolyzers are smaller than expected. Hydrophobicity
of the material is a serious problem as for other polymers. Improvements are needed
in this area.
46
Polyphenylene sulfide
Ryton, a polyphenylene sulfide thermoplastic resin, was developed by Phillips
Petroleum Company [57]. It has excellent thermal and chemical stability. No solvents
are known for the polymer below 185-205 °C. Ryton is rated for continuous use at
temperatures of 170-220 °C by Underwriters Laboratories.
Fibres of Ryton having a denier of 3-6 were melt spun by Phillips Fibres
Corporation. Fibres of 23 µm diameter were tested by Teledyne at 150 °C in KOH/O2
and KOH/H2 environments. The results of these tests appeared promising as the fibres
did not show any loss of tensile strength after a 500 h test period [51]. Fischer et al.
indicated that Ryton R6 is stable up to 200 °C in 50 % KOH and that it only corrodes
slowly in 70 % KOH at 250 °C [47].
Unfortunately, due to production problems, Ryton fibres are not presently
available. When these problems are resolved, testing of this product will be
warranted. Careful attention will have to be given to improving the wettability of the
fibres, as for the polysulphones.
Polytetrafluoroethylene
Polytetrafluoroethylene (P.T.F.E.) belongs to a family of polymers, the fluorocarbons,
having excellent chemical and heat resistance. P.T.F.E. has a maximum service
temperature of 288 °C and is highly resistant to alkaline media. It would be stable up
to 260 °C in 50 % KOH [47]. Vogt, however, has indicated that Teflon fibres could be
dissolved under certain conditions at 200 °C in 60 % KOH. It was suggested that the
large surface area of the fibres might favour the dissolution of the Teflon [28]. Teflon
is available as fibres, porous films or sheets.
One of the problems associated with the use of P.T.F.E. in any form as a
separator material is its lack of wettability. This favours formation of gas bubbles
within the separator and on its surface, thus increasing the contribution of ohmic
resistance to the cell voltage, and reducing gas purities. Surface treatments are,
therefore, necessary to make Teflon wettable. Grafting of stable anionic or cationic
groups onto Teflon could make the surface hydrophilic. However, it is generally
recognized that only anionic groups are stable in alkaline media at high temperatures.
Radiation grafting of acrylic acid groups on Teflon fabrics in the presence of a crosslinking agent is being studied by Sohm and Mas [58]. Some samples, treated at a low
grafting rate in the presence of a cross-linking agent, have shown lifetimes better than
2000 h in KOH at 200 °C. Results of life tests under electrolysis conditions at 120 °C
in 40 % KOH at an applied current density of 10 kA m-2 are not yet available.
According to Sohm and Mas, radiation grafting alone would not produce stable
anionic groups on the Teflon surface.
Another technique for improving wettability of porous Teflon is in-situ
formation of potassium titanates in porous Teflon sheets [59a]. Porous Teflon
membranes are impregnated with tetrabutyltitanate. The material is then hydrolyzed
in hot (60 °C) water for 3 h, dried at 350 °C, and calendared. A final treatment
consists of autoclaving the membrane at 200 °C in 30 % KOH for one hour [59b,c].
The result is the formation in the pores of whisker-shaped potassium titanate crystals.
Two types of porous Teflon sheet were tested: a collected-fibril type and an
expanded-sheet type. The expanded-sheet type did not retain the potassium titanate
crystals well. This new type of separator has better ionic conductivity than P.T.F.E.
which has been directly impregnated with Fybex potassium titanate; it has good
hydrophilic properties, and average pore sizes of 1- 2 µm. Ionic resistance values of
47
0.28 Ωcm2 and 0.7 Ωcm2 were obtained respectively for 0.4 mm and 0.7 mm thick
separators, in 30 % KOH at 25 °C [60].
The potassium-titanate impregnation technique appears most promising for
improving the wettability of Teflon membranes. Grafting techniques seem more
difficult to use and have yet to be proven for the electrolyser application.
Ion-exchange membranes
Ion-exchange membranes, anionic and cationic, are manufactured by a number of
companies including DuPont, Ionics, R.A.I. Research Corporation, Tokuyama Soda,
Asahi Chemical Industry and Asahi Glass. However, only a few of these membranes
can survive 150 °C temperatures in alkaline water electrolyzers. Cationic membranes
are the more resistant under these conditions; no anionic membranes are available
which are stable in alkaline media above 100 °C. This is due to instability of the
quaternary ammonium compounds used as the source of base; weaker bases could be
more stable but would lack conductivity.
Cationic ion-exchange membranes having a perfluorinated backbone and with
ion-exchange groups such as - SO3 or - COOH are the most stable under waterelectrolysis conditions. Unfortunately, it is not possible to obtain samples of the
promising perfluorocarboxylic-acid membranes which are fabricated by Asahi Glass
and Asahi Chemical Industry, their policy being to supply membranes only as part of
electrochemical
systems.
Other
promising
cationic
membranes
are
perfluorosulphonic-acid based membranes fabricated by DuPont and by R.A.I. [61].
Work at Brookhaven National Laboratories has shown Nation 115 and an unidentified
membrane from R.A.I. to be promising. A dependence of ionic conductivity of Nation
membranes on the type of electrolyte (KOH, NaOH, or LiOH), the caustic
concentration, the equivalent weight of the membrane, and membrane pre-treatment
was demonstrated by Yeo et al. [62]. In general, the ionic conductivity of the
membrane is dependent on its water content, which is related to the above parameters,
the conductivity being maximum for a high water content. Membrane conductivity
would be optimum for 1000 equivalent-weight membranes in NaOH (or LiOH)
electrolyte at approximately 10 % concentration, thin membranes and high
temperatures, and also for membranes swollen at high temperatures [62]. Nafion has
generally excellent chemical and physical stability in alkaline media. Neutralization
of the sulphonic-acid groups increases the glass-transition temperature from 110 to
220 °C [62]. However, use might be limited to "low" alkali concentration; Fischer
reported that Nation is attacked severely in 50 % KOH above 150 °C [47]. No
detailed results are available for tests on R.A.I. membranes at high temperature.
Perfluorinated cationic ion-exchange membranes could provide a promising
replacement for asbestos in high-temperature water electrolyzers. Tests are necessary
to confirm their stability in alkaline media at 150 °C, and also to determine their
resistance to ionic conduction under the same conditions. Membranes with ionexchange properties produced by polymerization in situ of organic compounds on
asbestos fibres or cardboards or other porous supports are being investigated. These
membranes would be more economical than Nation. No long-term results are
available [63,64].
Polybenzimidazoles
Polybenzimidazoles (PBI) are complex polymers, based on aromatic, nitrogencontaining rings. Many of the imidazole derivatives are resistant to the most drastic
48
conditions. They are not readily attacked by oxidizing agents and have high melting
points and excellent stabilities at elevated temperatures (430 – 645 °C) [65].
Polybenzimidazole fibres, prepared by Celanese Corporation, were tested at 121
°C in KOH for 5000 h and showed a weight loss of only 1 % [36]. No changes in
fibre diameter or morphology were detectable. Work by Chenevey suggests that this
material might be stable up to 100 °C [66].
Tensile-strength measurements made under the Noranda/Electrolyser program
on two types of PBI cloth after exposure to 30 % KOH at 80 °C showed the material
to lose 80 % of its strength after one month's exposure and up to 95 % after 15
months; the same PBI cloths showed 10-15 % shrinkage. In view of these results, it is
doubtful that PBI fibres could survive the alkaline-water-electrolysis environment at
high temperatures.
Other polymers
Polyquinoxaline,
polyphenylquinoxaline.
Polyquinoxaline
(P.Q.)
and
polyphenylquinoxaline (P.P.Q.) are polymers having aromatic, nitrogen-containing
rings. These polymers have excellent oxidative and thermal stability. Polymerdecomposition temperatures for some phenyl-substituted polyquinoxalines are as high
as 550 °C [67]. P.P.Q. showed excellent chemical and dimensional stability in 45 %
KOH at 80 °C after an extended period of exposure [68].
Polymer H or H-resins. H-resins are thermosetting polyphenylene polymers
developed by Hercules Incorporated. The material has very good resistance to severe
chemical and thermal environments. Tests indicated good stability in KOH at 190-200
°C. Fibres with diameters of 5-10 µm were fabricated [69], but work on the
development of this resin has now been stopped by Hercules.
Electrodes
There are no generally valid rules for the shape of the electrodes. Since the internal
resistance of an electrolytic cell should be as small as possible in order to keep the
expenditure of energy as low as possible, the general attempt is to make the electrode
surfaces as large as possible and the electrode spacing as small as possible. Large
surfaces are achieved by using flat plate construction or by using large-area sheets or
strips.
To increase the surface area of the plates, these are in many cases additionally
roughened by sandblasting. With particularly expensive electrode materials (e.g.,
platinum), thin foils, perforated foils or grids or wires are wrapped around plates of
insulation material (glass etc.).
In achieving the smallest possible distance between the electrodes, it must be
ensured that that minimum value is not exceeded or the occurrence of an impermissibly high bath intermixing will again impede the current yield, and thus reduce
the efficiency of the cell. The electrode spacing determined during the design of a cell
is obtained by the use of fixed, or in some cases adjustable spacers made of glass or
cement, so that the electrodes do not lie directly against the diaphragms.
The choice of the electrode materials is governed by a number of aspects. The
requirements of a good electrode material are: maximum electrical conductance, high
corrosion resistance and minimum overvoltage. Selection of the electrode material is
also influenced by the proposed electrolyte. In conventional water electrolysis, the
cathode is usually mild steel and the anode, which is subject to greater corrosion, is
almost exclusively of nickel. Nickel has traditionally been used as the anode material
in alkaline water electrolysis. It is highly corrosion-resistant at positive potentials in
49
alkaline electrolytes. The oxygen evolution efficiency of nickel is among the highest
for elemental metals [70].
Because of the high price of nickel, the anodes are not, as a rule, manufactured
of solid nickel, but only provided with an electrolytically deposited nonporous nickel
coating.
Four main factors must be considered in developing a commercially practical
anode for alkaline water electrolysis: electrochemical efficiency, stability, scale-up,
and cost. Efficiency is the initial screening criterion in any anode development
program. Coupled with the need for high electrocatalytic activity is the requirement of
low internal anode resistance. This resistance depends on both the anode materials
and structure. Once high efficiency has been demonstrated, the anode must be tested
for stability. As a rough goal, physical and chemical degradation should be minimal
during several, even as many as ten or more [71], years of operation. The anode must
not only be stable at oxygen evolving potentials, but must resist open-circuit corrosion
as well. With the recent advent of highly efficient cathode catalysts, the anode
corrosion products, however limited, must not foul the cathode surface. In addition to
chemical and physical stability, the anode should provide a constant potential after an
initial break-in period, rather than the time-dependent potential increase which occurs
at nickel anodes [72- 74]. Manufacturing scale-up is a necessary but difficult part of
commercial anode development, involving a move from the laboratory to a pilot
processing facility. A way must be found to manufacture anodes on the order of 1 m s
in area. The process must be under sufficient control to assure constant anode quality.
Because of cost constraints, an automated or semi automated manufacturing process is
necessary.
The maximum practical anode cost depends on the capital cost constraints for
the entire electrolysis system. This, in turn, depends not only on competing
electrolysis systems, but on non-electrolytic hydrogen production alternatives as well.
The permissible cost of any anode improvement will be based on the power savings
realized from it and the anticipated anode lifetime.
There are several approaches to improving the efficiency of alkaline water
electrolysis. Raising the electrolysis temperature, for example, lowers the voltage
required to maintain a given cell current density [75]. Improvements in cell design,
separator structures, and materials, etc., will also contribute to better cell performance.
Whatever advances are made, it will remain necessary to develop electrodes,
compatible with the system design and operating conditions, which will give the
lowest possible overpotentials. This can be achieved by two methods, which can
sometimes be combined for maximum benefit. In the first, catalytically active
materials such as NiCo2O4 [76] are applied to the electrode surface. The second
method involves greatly increasing the electrode surface area (i.e., its "roughness
factor," defined as the ratio of its real surface area to its apparent, or geometric, area),
thereby lowering the real current density and the associated activation overpotential.
Nickel anode efficiency has been improved by developing high surface-area anode
structures. For example, porous, high surface-area anodes have been made by
sintering fine nickel powders prepared by nickel tetracarbonyl decomposition. These
anodes, although generally less porous, are essentially similar to the positive plaques
used in nickel-cadmium batteries. In high-pressure (30 atm) electrolysis at 200 oC the
oxygen evolution overpotential, ηO2, on sintered, porous nickel anodes was about 100
mV lower than on smooth nickel anodes, at current densities of 500-1500 mA/cm2 in
35 % KOH electrolyte [77]. However, sintered, porous nickel anodes were only
marginally more efficient than nickel cloth anodes in atmospheric pressure
7
50
electrolysis, over the temperature range 25 o-90 oC [81]. This was attributed to poor
inner surface utilization in the porous anode at atmospheric pressure, presumably due
to gas blockage. Similar results were obtained in 5N KOH at 25 oC, i.e., 100 mesh
nickel screen and sintered, porous nickel anodes produced approximately equal
current densities when compared potentiostatically at 1.64 V/DHE [78].
Efficiency gains, as a rule, must be balanced against the additional capital cost
incurred [79]. For this reason, work continues to be done to minimize oxygen and
hydrogen overpotentials on inexpensive electrode materials such as nickel [80] and
nickel alloys [81,82], which are the least expensive electrode materials currently
known that can be used as both anodes and cathodes. These materials function most
efficiently when fabricated as high surface area electrode forms.
The electrodes prepared by applying particulate metal coatings to metal
substrates. The use of a paint-like suspension of the metal particles makes it possible
to achieve uniform coatings of controlled thickness using a variety of coating
techniques. This electrode fabrication method, which appears to be well suited for
preparing anodes for alkaline water electrolysis, has two main advantages:
(i) the coating technique can be used to add considerable surface roughness to
existing electrode structures of many kinds;
(ii) the materials and procedures used to prepare the electrodes are relatively
inexpensive.
Both of these factors are in keeping with the capital cost restraints on
commercial alkaline electrolysis equipment.
The efficiencies of the coated electrodes and their microstructures can be varied
significantly by changing the coating preparation conditions. The interior surfaces of
both anode and cathode coatings participate in the electrode reactions to some extent;
thus, overpotentials can be reduced by increasing the electrode coating thickness.
However, the use of thick anode coatings to improve electrolysis efficiency is of
doubtful value on economic grounds, since the overpotential reduction obtained is
small. This would be true even under ideal conditions, in which the overpotential
reduction for a tenfold increase in coating thickness is given by the Tafel slope, which
is only ~35 mV [83].
Anodes with thin, sintered, porous nickel coatings were made by applying nickel
powders, in polysilicate-based slurries, to nickel and steel substrates [83]. The
coatings were sintered under a reducing atmosphere. The nickel coating morphology
depended on the sintering temperature and, to a lesser extent, the sintering time.
Decreasing the sintering temperature through the range 980 o-760 o C and the
sintering time through the range 30-5 min progressively preserved more of the fine
structure in the coatings. A corresponding ηO2 decrease of about 70 mV was observed
in electrolysis at 200 mA/cm2 in 30% KOH electrolyte at 80 oC. Although the interior
coating surfaces participated in the oxygen evolution reaction to some extent, the ηO2
reduction obtained by increasing the coating thickness was small. In a subsequent
study [84], the porous nickel-coated mild steel anodes were operated for 1200h at 100
mA/cm2 in 30% KOH at 80oC. After an initial rise, the anode potential was essentially
constant. Nickel-iron alloy formation in the steel surface region, during the high
temperature sintering treatment used to bond the coating to the substrate, greatly
increased its corrosion resistance. Nickel coatings made by slurry coating and
sintering nickel powder have also been applied to foamed nickel anode supports [85].
The resulting anodes were operated at 500 mA/cm2 in 40% KOH electrolyte at 90oC.
The voltages of cells with porous nickel/foamed nickel anodes and cathodes were
about 50-100 mV lower than those of cells with uncoated foamed nickel electrodes.
51
Cells equipped with the coated electrodes maintained a stable 1.68V operating voltage
at 200 mA/cm2 in 100h electrolyses at 118oC. The results cited above show that
coatings made by sintering fine nickel powder prepared by Ni carbonyl
decomposition reduce oxygen overpotential. Avoiding the high temperatures used for
nickel sintering, which produce a concomitant loss in surface area, would appear to
offer the advantage of maximized anode surface area. However, Balajka found that
TFE-bonded nickel powder applied to 100 mesh nickel screen was ineffective [86].
Such anodes were actually about 60 mV less efficient than similar screens with
electroplated nickel coatings, at a current density of 370 mA/cm2 in 30% KOH
electrolyte at 80 oC. This poor performance was ascribed to possible oxidation of the
metal surface during electrode preparation.
Nickel whisker anodes [87,88] were made from polycrystalline nickel whiskers
grown by chemical vapour deposition of Ni(CO)4 gas in an electromagnetic field.
Whiskers, with diameters from 0.1 to 50 µm and lengths from one-tenth to several
centimetres, were sintered at 800-1000 oC into continuous, fibrous networks on 200
mesh nickel screen. The resulting anodes were up to 90% porous, with specific
surface areas of up to 5 × 103 cm2/g. The anodes were tested in 30% KOH electrolyte,
at current densities from 100 to 1000 mA/cm2, based on apparent anode surface area.
Oxygen evolution overpotentials were about 100 mV lower on whisker anodes than
on multilayer, 200 mesh nickel screens. There was no loss of structural integrity
during 48h of electrolysis at 1000 mA/cm2.
Although the electrode structure is less suitable for hydrogen evolution than for
oxygen evolution, the cathode overpotential reduction with increased coating
thickness is larger. This is not due to the more effective utilization of the electrode
coating interior for hydrogen evolution than for oxygen evolution, but is rather a
consequence of the difference in Tafel slopes for the two processes.
The optimum electrode coating microstructure, for any given cell and set of
operating conditions, is that which gives adequate strength and stability while
maintaining the highest possible surface roughness factor. The different coating
strengths are necessary, depending on whether the coated electrodes are used as
anodes or cathodes. The efficiency improvements which can result from choosing
proper sintering conditions are large enough to be significant. In addition, sintering no
more than necessary for adequate mechanical integrity of the coating minimizes
sintering costs.
The use of steel substrates with porous coatings as anodes in alkaline
electrolysis raises questions regarding their durability, since unprotected mild steel
corrodes rapidly under anodic service. None of the electrodes used showed evidence
of corrosive failure in electrolyses.
Efforts to improve the efficiencies of nickel anodes have focused on raising
their effective surface areas [83,89] as well as applying electrocatalysts to their
surfaces [90].
Raney nickel is made by alloying nickel with metals such as aluminium or zinc.
When the Raney alloy is leached in alkaline electrolyte, a high surface-area structure
with high electrochemical activity is produced. Several approaches to making Raney
nickel anodes for alkaline water electrolysis have been reported. Raney Ni-Zn alloy
was prepared by electrodeposition from chloride electrolyte [48]. The
electrodeposition conditions affected both the composition and structure of the alloy.
A maximum BET specific surface area, greater than 25 m2/g, for the leached anodes
were obtained at a deposition potential of 1.07 V/SCE. Correspondingly, electrolysis
cells containing Raney nickel anodes and cathodes had minimum voltages using alloy
52
electroplated at -1.07 V/SCE. Analysis of the experimental data indicated that the
improved performance of Raney nickel-coated anodes was due solely to greater anode
surface area [91]. An electrolysis cell containing Raney nickel-activated iron anodes
and cathodes was stable for 3000 h at a current density of 200 mA/cm2 and a
temperature of 110 oC.
Electrodeposition of Ni-Zn alloy onto nickel or stainless mesh has also been
used to make Raney-type porous electrodes by plating for as much as 24 h [92]. These
electrodes differed from the usual sintered powder porous anodes in their pore
structure, which was channel-like and full of very fine hairline cracks. Porosities
ranged from 65% to 75%, average pore sizes from 1 to 3 µm. Electrochemical
measurements in 6N KOH showed that the electrodeposited porous anodes were
generally inferior to sintered ones.
A plasma-sprayed Ni-A1 anode coating produced an initial overpotential
reduction of 200 mV in electrolyses at 160 o and 200 oC [93]. After 100h, the ηO2
improvement was still 180 mV at 160 oC whereas at 200 oC the improvement
decreased to 150 mV after 350 h. In contrast, Raney nickel prepared by plasma
spraying Ni-A1 alloy particles was found to be a poor oxygen evolution catalyst in
another study [94]. This was attributed to rapid oxidation of the anode surface. The
Raney nickel structure was found, however, to be useful as a support for other
catalysts.
Raney nickel deposited on foamed nickel anode supports behaved similarly to
sintered, porous nickel coatings on foamed nickel, i.e., cell voltages were 50-100 mV
lower when Raney nickel-coated anodes and cathodes were used instead of uncoated
foamed nickel electrodes85. In 40% KOH electrolyte at 118 oC Raney nickel was
initially superior to sintered, porous nickel, but the advantage declined to only 20 mV
within 100 h.
Various Raney alloy anodes tested at 90 oC were about 60 mV more efficient
than conventional, low surface-area nickel anodes [95]. At a current density of 100
mA/cm2, ηO2 was 240 mV for Raney nickel, 240 mV for Raney nickel-cobalt (30
atom percent (a/o) Co), and 230-240 mV for Raney cobalt. These values compared
with 300 mV for steel-blasted nickel and 280 mV for electrodeposited nickel.
Several methods of doping Raney nickel anodes with lithium were studied by Martin
et al. [96]. In almost all cases, anode activity declined greatly during 24 h of
electrolysis at 80 oC at a current density of 1000 mA/cm2. Best results were obtained
by pre-oxidizing the anodes in H2O2 + LiOH or thermally in air. In the same study,
better results were obtained when Raney nickel anodes were impregnated with cobalt
oxide. Overpotential reductions of more than 130 mV were measured after 24h of
electrolysis at 80 oC at a current density of 1000 mA/cm2. Impregnating Raney nickel
with mixed cobalt-nickel oxides was no more effective than using cobalt alone.
Nickel anode degradation during high temperature service has been
investigated, primarily in conjunction with the French high-temperature electrolysis
programs. Giles [97] analyzed the physical and electrical characteristics of rollcompacted, sintered, porous nickel anodes before and after pilot plant operation at
CEM. Electrolysis was carried out at 1000 mA/cm2, in 35% KOH at 200oC under 30
atm pressure. During 250 h of electrolysis, anode porosity decreased from about 45%
to about 20% as corrosion products accumulated within the anode. As the finely
porous corrosion products filled the larger anode pores, the mean pore diameter fell
from 1.9 to 0.3 µm after 23 h and to 0.1 µm after 83 h. X-ray diffraction showed both
Ni(OH)2 and NiO, but no +3 nickel species, in the corrosion product. Anode tensile
strength dropped from 50 to 20 N/mm2 after 83 h of electrolysis. The electrical
53
resistivity increased from about 27 × 10-6 Ωcm to about 46 ×10-6 Ωcm during the
same period. Despite the slowing of the nickel corrosion rate with time because of the
partly protective nature of the corrosion layer, Giles concluded that using sintered,
porous nickel anodes in high-pressure, high-temperature service was questionable.
A joint GDF/EDF study [98] addressed nickel corrosion in the temperature
range 130o- 180 oC. While positive potentials were not imposed on the nickel
specimens, the corrosion tests were carried out in an oxygenated environment in 40%
KOH. Specimens of Ni 200 and Ni 201 industrial castings, a powder metallurgy
compact (Ni 270), and high-purity, zone-refined nickel were studied. In the
oxygenated electrolyte, the porous corrosion products were NiO and Ni(OH)2 at 130
o
C while only NiO was observed at 180 oC. Corrosion rates at 180 oC were 10-15
µm/yr for high-purity nickel, and 30-45 µm/yr for Ni 200, 201, and 270. Restricting
the sulphur content of the nickel was found to retard corrosion.
Under anodic polarization, the Ni+2 species can be converted to Ni+3.
Electrochemical and ellipsometric studies [74] demonstrated that trivalent nickel was
the desired species for efficient oxygen evolution. Thus, Ni(OH)2 may be regarded as
the electrocatalyst precursor. Attempts to catalyze nickel anodes with Ni(OH)2 have
benefited from the extensive work on nickel electrodes for alkaline storage batteries
[99]. Sintered, porous nickel anodes impregnated with Ni(OH)2, essentially nickel
battery positives, were considerably more efficient for oxygen evolution than similar
unimpregnated anodes [80]. Electrolysis was carried out at 90oC in 34% KOH
electrolyte. At a current density of 400 mA/cm2, ηO2 was only about 100 mV on the
impregnated anodes, while the unimpregnated anodes were, as discussed earlier, no
more efficient than nickel cloth anodes. This enhanced activity was attributed to
higher superficial surface area rather than improved utilization of the inner anode
surfaces. Anodes impregnated with 1.91g of Ni(OH)2 per cm3 of void volume, a
characteristic nickel battery positive loading, showed rapid disintegration under
continuous charging at 80oC This was presumably due to high internal stresses.
Anodes with a lower Ni(OH)2, loading of 1.15g/cm3 of void volume were
dimensionally and structurally stable during more than 100 days of operation at 400
mA/cm2.
In a study of nickel battery positive plates, the oxygen evolution reaction was
observed at an earlier stage of charging when the plates were impregnated
electrochemically rather than chemically [100]. Water electrolysis anodes in sheet and
woven screen form, with sintered, porous nickel coatings, were electrochemically
impregnated with Ni(OH)2 [101]. In a one-step impregnation process, the amount of
Ni(OH)2 precipitated varied linearly with the charge passed. This allowed the catalyst
loading to be controlled accurately. On Ni(OH)2-impregnated anodes, ηO2 at 200
mA/cm2 was 45-60 mV lower than on similar un-catalyzed anodes, in 30% KOH
electrolyte at 80oC. Optimum Ni(OH)2 loadings were about 1-4 mg/cm2 of apparent
anode surface area. At higher loadings, overpotentials rose as the outer pores of the
nickel coatings became plugged with Ni(OH)2, resulting in a loss of effective anode
surface area.
Lithiated NiO electrocatalyst was prepared by vacuum decomposition of a
mixed slurry of Ni(OH)2 and LiOH [102,103]. Lithium, present at 1 a/o in the
finished catalyst, greatly reduced the electronic resistivity from the un-doped value
(>108 Ωcm) to 100 Ωcm. The BET specific surface area of the catalyst was 100 m2/g.
Anodes were made by mixing the catalyst with TFE binder and sintering onto 100
mesh nickel screen. In 5N KOH electrolyte at 60 oC the catalyzed anodes were about
54
55 mV more efficient than un-catalyzed nickel screens and were comparable to
NiCo2O4-catalyzed anodes.
More fundamental studies have investigated the oxygen evolution reaction
mechanism [104] and the nature of the surface on which the reaction takes place
[105].
The oxygen evolution characteristics of the mixed oxide spinel, NiCo2O4, have been
reported extensively, most notably by Tseung and co-workers [76,102,106- 109]. In
strongly alkaline electrolytes at high positive potentials, i.e., conditions typical of
commercial electrolysis, oxygen evolution proceeds by decomposition of species with
tetravalent cations [108,109].
NiCo2O4 has been made in several ways. Freeze-drying mixed Ni and Co
nitrates, followed by vacuum decomposition at 250 oC and a final heat-treatment for
10h at 400 oC in air, is preferred. The resulting catalyst has a high specific surface
area (70 m2/g by BET) and a moderate electrical resistivity (10 Ωcm) [102]. Anodes
made with freeze-dried NiCo2O4 were more efficient than similar anodes catalyzed
with NiCo2O4 made by mixed nitrate co-precipitation [107]. It was proposed that the
freeze-dried product was more homogeneous, i.e., there was less separation of Ni and
Co oxides. NiCo2O4 has also been made by direct thermal decomposition of the mixed
nitrate salts on anode substrates102. However, the anodes were less efficient than those
made with freeze-dried NiCo2O4.
Highest anode efficiencies have been obtained with the freeze-dried NiCo2O4
incorporated into a TFE-bonded coating [102]. Optimum efficiency was obtained
using 15%-30% TFE binder. For example, at 60 oC in 5N KOH, a catalyst loading of
13 mg/cm2 produced a current density of 1000 mA/cm2 at 1.6 V/DHE. However,
oxygen bubbles were observed to cling to the hydrophobic anode surface. Applying a
thin, hydrophilic nickel cobalt oxide layer to the anode surface improved bubble
release. This reduced the anode potential by about 20 mV at 1000 mA/cm2 in 5N
KOH at 25oC. In another study [86], TFE-bonded NiCo2O4 anodes were about 150
mV more efficient than nickel-plated nickel screen when operated at a current density
of 370 mA/cm2 in 30% KOH electrolyte at 80oC.
Doping NiCo2O4 with varying amounts of Sr and La was effective in reducing
ηO2 [96]. Best results were obtained when 0.5 atoms of (2/3 Sr + 1/3 La) per atom of
cobalt were added. The resulting material, with nominal composition
NiCo2Sr0.66La00.33O5.17, produced 300 mV overpotential at a current density of 1000
mA/cm2.
NiCo2O4-catalyzed anodes have shown high oxygen evolution efficiency in
several laboratories. Long-term anode stability, however, is open to some question. At
a current density of 330 mA/cm2, an overpotential decrease of 150 mV relative to
nickel anodes was reported [72]. However, poor stability was noted at temperatures
above 100oC and current densities greater than 200 mA/cm2. In contrast, thermally
decomposed NiCo2O4 anodes were stable for 2000 h of electrolysis at 1000 mA/cm2
at a temperature of 120oC. In another study, NiCo2O4 anodes were stable for over
1000h of electrolysis at 1000 mA/cm2, at temperatures from 120oC to 200oC [26].
However, the catalyst was very sensitive to corrosion at open circuit.
In tests of about 150 days duration at low current density (75 mA/cm2) and
temperature (70 oC in 28% KOH, a TFE-bonded, freeze-dried NiCo2O4 anode was
initially about 60 mV more efficient than a nickel-plated steel anode [73]. After 150
days, due to a lower ηO2 increase with time, the NiCo2O4 anode advantage was over
1 1
55
100 mV. Similarly, in 18 days of electrolysis at 330 mA/cm2, ηO2 was about 100 mV
lower on NiCo2O4 anodes than on un-catalyzed Ni screen in 30% KOH at 90 oC [27].
The spinel Co3O4 has shown promising efficiency and long-term performance.
Thermally decomposed Co3O4 was deposited on electroformed nickel plates with
conic holes [110] and was tested in atmospheric pressure electrolysis at 120oC. After
6000h of electrolysis at 1000 mA/cm2 in 40% NaOH electrolyte, the anode potential
was about 0.59V vs. Hg/HgO reference (containing the bulk electrolyte composition).
After an initial break-in period, the rate of ηO2 rise with time was only 4 mV/1000h.
Co3O4 anodes containing 0, 4, 7, and 10 a/o Li dopant were prepared by freeze
drying and vacuum decomposing the mixed Li and Co nitrates [106,111]. This was
followed by heat-treatment at 600 oC for 10h to produce the Li-doped spinels. The
electrical resistivities of the catalyst powders varied greatly with the extent of Li
doping, from 104 Ωcm at 0% Li to 10 Ωcm at 4% Li, and 1 Ωcm at both 7% and 10%
Li. TFE-bonded anodes were made using 10/3 weight ratios of the catalyst powders to
TFE.
These anodes were operated at 1000 mA/cm2 in 5M KOH at 70 oC. The anode
potential dropped markedly with increasing Li doping, i.e., 1.69V with no Li dopant,
and 1.56, 1.535, and 1.52V, respectively, at 4, 7, and 10 a/o doping. The enhanced 02
evolution activity with increased Li doping was attributed to the increased fraction of
Co +3 ions which could form if all the Li entered tetrahedral lattice sites [111].
An electrolysis cell equipped with a TFE-bonded, 10% Li-doped Co3O4 anode
was operated for about 6000h at 1000 mA/cm2 [106]. The cell contained 45% KOH
electrolyte at 85 oC. An anode half-cell measurement after approximately 5600 h
indicated an ηO2 increase of about 50 mV during operation. The Li content of the
anode was unchanged. A TFE-bonded, Li-doped Co3O4 coating was applied to a
porous nickel anode substrate [77]. In 30% KOH electrolyte at 70 oC at a current
density of 1000 mA/cm2, ηO2 was 300 mV at the beginning of electrolysis and 296
mV after 1006 h of operation.
A variety of AB2O4 spinel compounds (B = A1, Cr, Mn, Fe, or Co) were used
to prepare TFE-bonded anodes [60]. The anodes were tested in 30% KOH at 80 oC.
At a current density of 1000 mA/cm2, the spinel anode overvoltages were 30-70 mV
lower than the overvoltage of un-catalyzed nickel.
The mixed oxides NiLa2O4, NiPr2O4, and NiNd2O4 were prepared by thermal
decomposition on platinum supports [112]. In potentiostatic experiments conducted in
30% KOH at 50oC the three catalysts showed similar current vs. potential
characteristics. A low Tafel slope (approximately 40 mV/decade) was observed at
current densities of up to 100 mA/cm2. At higher current densities, the Tafel slope
was about 120 mV/decade. The anode potentials were approximately 1.6 V/RHE at
1000 mA/cm2. The oxygen evolution overpotentials on the mixed oxide anodes were
observed to be essentially stable for several hundred hours, at current densities as high
as 1000 mA/cm2.
The NiLa2O4 anode electrocatalyst has been doped with Li, Mg, and Fe,
replacing some of the Ni, and with partial substitution of sulphur for oxygen [78]. The
surface properties of the Li-doped materials were claimed to be slightly superior to
un-doped NiLa2O4. However, Li doping produced an increase in bulk catalyst
resistivity as x in Ni1-xLixLa2O4 was increased from 0 to 0.24, in contrast to the lower
resistivity produced by doping Co3O4 with Li, as discussed in the preceding section.
The net effect on ηO2 was negligible. Mg doping also increased the electrical
resistivity of the catalyst. Similarly, high levels of Fe doping, e.g., x = 0.06 in Ni1-
56
xFe2x/3La2O4,
increased electrical resistivity. Lower Fe doping (e.g., x = 0.03)
produced a resistivity decrease, but failed to improve ηO2 for the NiLa2O4 anode.
Cobalt substitution, while effective in reducing ηO2, produced a non-stoichiometric
oxide of nominal formula Ni0.8Co0.2LaO3. Its anodic polarization was about 25 mV
lower than that of NiLa2O4 in 30% KOH electrolyte at 85 oC over the current-density
range 100- 1000 mA/cm2. At the current density of 500 mA/cm2, the Co containing
anode also showed superior time behaviour during 300h of electrolysis, producing a
~100 mV lower anode potential during the last 150h. However, raising the electrolysis
temperature to 110 oC produced a significant deterioration in performance, so that the
Co containing anode became inferior to NiLa2O4. Treatment of NiLa2O4 with H2S to
incorporate sulphur into the catalyst worsened the anode polarization behaviour. This
was attributed to the higher resistance of the sulphur containing anodes.
Perovskite anode catalysts of the La1-xSrxCoO3 type, and substitutional
variants, have been studied more extensively than any other mixed oxides, with the
possible exception of NiCo2O4. Their high activities for oxygen evolution are similar
to that of NiCo2O4, with the relative superiority of the two types of catalyst still open
to question. For example, in 30% KOH electrolyte at 145 oC, La0.5Sr0.5CoO3 was
about 120 mV superior to NiCo2O4 at 1000 mA/cm2 [26]. However, under similar
conditions (120 oC 1000 mA/cm2), cells with La0.5Sr0.5CoO3 anodes showed 0.1V
higher voltages than similar cells with NiCo2O4 anodes [113]. A mixed La-Sr-Co
oxide of unspecified composition, deposited on a Raney nickel substrate, was 180 mV
more efficient than an un-activated nickel mesh anode at a current density of 1000
mA/cm2 in 40% KOH electrolyte at 160 oC [94]. The anode was stable for more than
1000h.
It is generally agreed that the La-Sr-Co perovskite composition has a
significant effect on oxygen evolution overpotential. The relationship between ηO2
and x in the formula
La1-xSrxCoO3is not completely clear, but an overall trend
toward higher catalytic activity in the compositional midrange has been determined.
When 5-10 mg/cm2 La1-xSrxCoO3 coatings were applied by spray pyrolysis, a broad
ηO2 minimum, centred at about x = 0.5, was observed at a current density of 10
mA/cm2 [114]. At a current density of 500 mA/cm2, optimum overpotentials of about
330 mV were measured over the approximate range 0.5 < x < 1.0. These results were
obtained in 6N KOH electrolyte at 80 oC. In a later study with more data points
reported, the variation in ηO2 with x was refined [96]. At current densities of 100,
500, and 1000 mA/cm2, an overpotential minimum was found at about x = 0.75, while
x 0.15 produced an overpotential maximum. At low current densities in 45% KOH at
25 oC La0.5Sr0.5CoO3 and La0.8Sr0.2CoO3 had approximately equal potentials, while
both were superior to LaCoO3 [115].
The variation in electrical resistivity, ρ, vs. x for several perovskite families
La1-xSrxCoO3 is roughly similar to that described above for ηO2 vs. x [116]. This
suggests that electrical resistivity is an important factor in the efficiency of perovskite
anode catalysts. For most of the lanthanides studied, ρ was 0.1-10 Ωcm at x = 0,
decreasing to about 10-3.5 Ωcm at x = 0.5. The minimum ρ composition interval for La
containing perovskites was quite broad, from about x = 0.2 through x = 0.6, the
highest value investigated. The ρ decrease as Sr was substituted for La has been
attributed to more Co+3/Co+4 couples in the doped perovskite [115].
La0.5Sr0.5CoO3 coatings have been prepared by a number of methods. Spray
pyrolysis was used to deposit 7 mA/cm2 coatings in a study of the effects of final
calcination temperature [93]. Best results were obtained by calcining the coatings at
600 o – 700 oC. In 8N KOH at 80 oC was about 380 mV at 1000 mA/cm2. Raising the
57
coating loading to about 15 mA/cm2 further reduced ηO2 to about 340 mV at 1000
mA/cm2. Thermally decomposed coatings, containing 1.2 mA/cm2 La0.5Sr0.5CoO3,
were also studied [95]. At 90 oC these coatings gave ηO2 approximately 250 mV at
100 mA/cm2 and 330 mV at 1000 mA/cm2. From ηO2 at 90 oC and the rate of
overpotential change with temperature, d ηO2 /dT, an overpotential of 220 mV was
estimated for 1000 mA/cm2 electrolysis at 160 oC. TFE-bonded anodes (36% TFE, 14
mA/cm2 loading) were nearly as active as anodes without TFE at low current
densities, in 6N KOH at 80 oC [114]. However, ηO2 increased rapidly at current
densities above 10 mA/cm2 due to resistance losses between catalyst particles in the
coating. Changes in TFE content and preparation method did not overcome this
problem.
The effects of electrolysis temperature on the activity of La1-xSrxCoO3 are
unclear. It is generally agreed that, at temperatures up to 160 oC an increase in
electrolysis temperature produces an increase in catalyst activity. Because d ηO2 /dT
for the perovskite catalysts is greater than for other anodes (e.g., Raney nickel, unactivated nickel) their relative improvement is greater with increasing temperature
[95]. Continuous improvement up to 200 oC was reported for electrolysis in 36%
KOH at 22 atm pressure [114]. The ηO2 improvement at 1000 mA/cm2 relative to unactivated nickel anodes was 260 mV at 160 oC and 280 mV at 200 oC. However, a
subsequent publication from the same laboratories [93] reported a decrease in activity
between 160 oC and 200 oC at 1000 mA/cm2 in 8N KOH. The experimental evidence
indicated that this was an intrinsic property of the catalyst.
The stability of La-Sr-Co oxides under anodic polarization has been good. In
36% KOH at 160 oC La0.5Sr0.5CoO3 was stable during 500h of electrolysis at 1000
mA/cm2, producing a 150 mV ηO2 reduction [93]. A series of La1-xSrxCoO3
compounds, prepared by freeze drying, were corrosion resistant at 220 oC in 75%
KOH electrolyte [115]. No activity loss was observed during 1000h of continuous
operation, and 2600h of operation with periodic interruption, at 1000 mA/cm2 in the
temperature range 120 o -200 oC [26]. However, La-Sr-Co oxides were reported to be
very sensitive to passive corrosion.
Several anode electrocatalysts have been made by substituting nickel or iron
for some or all of the cobalt in La1-xSrxCoO3. The oxygen evolution activity in the
family of compounds La1-xSrxFe1-yCoyO3 increased with increasing x and y [117].
SrCoO3 itself could not be synthesized as a perovskite. As a result, the most efficient
anode - catalyst prepared was La0.2Sr0.8Fe0.2Co0.8O3. In 1M KOH electrolyte at 25 oC
that catalyst produced a steady potential of 0.90V vs. Hg/HgO at 100 mA/cm2, after
an initial potential rise during the first 40h of electrolysis.
Nickel substitution for cobalt, in the series of compounds La1-xSrxFe1-yCoyO3,
has also been investigated [93]. Coatings with 10 mA/cm2 of the perovskites were
prepared by spray pyrolysis and tested in 8N KOH at 80 oC. At a current density of
1000 mA/cm2, a minimum ηO2 of about 310 mV was obtained at about y = 0.6. The
compound LaNi0.2Co0.8O3 was plasma sprayed to produce 8 mA/cm2 coatings [95]. At
90 oC ηO2 was 270 mV at 100 mA/cm2 and 330 mV at 1000 mA/cm2. The 90 oC data
and d ηO2/dT were used to estimate an overpotential of 240 mV at 160 oC at a current
density of 1000 mA/cm2. An anode catalyzed with La0.2Sr0.3Ni0.4Co0.6O3 produced an
initial 140 mV overpotential reduction at 200 oC but lost activity after 1h of
electrolysis [96]. Surface analysis showed an increase in the Sr/La ratio.
The perovskite electrocatalyst LaNiO3 was prepared by co-precipitation of La
and Ni nitrates, followed by an oxidizing heat-treatment at 800 oC [118]. Anodes
were made by pressing the LaNiO3 powder into pellets and sintering at 750 oC. In
58
approximately 1M hydroxide electrolytes at 25 oC the rate of O2 evolution on LaNiO3
was about 105 times faster than on Pt and about 102 times faster than on NiCo2O4. At
a current density of 100 mA/cm2, ηO2 was about 310 mV. The Tafel slope was 40
mV/decade. The LaNiO3 surface was non-stoichiometric, with almost all the nickel in
the +2 state. It was proposed that the high activity of LaNiO3 was due to its high
surface carrier density (approximately 1020/cm3).
Bockris and co-workers suggested a course for future perovskite
electrocatalyst development [119]. Their XPS examinations indicated a
nonstoichiometric surface, in which charge is mainly compensated by oxygen
vacancies. Because of the strong affinity of oxygen-deficient surfaces toward oxygen
in the form of OH, an increase in electrocatalytic activity would be expected with
increasing Mz-OH bond strength, where Mz is a surface transition metal ion. As a
result, it was proposed that perovskite electrocatalyst research should focus on finding
a bonding sequence which produces low intermediate radical coverage. In an
experimental study covering 18 perovskite compositions, carried out by the same
group, the electrochemical desorption of OH was found to be rate determining in
every case [120]. Later the study of perovskite electrocatalysis was extended [121]. It
was shown that reaction rates increase with decreases of magnetic moment, stability
of the perovskite lattice, and enthalpy of transition metal hydroxide formation, and
with increased number of d-electrons in the transition metal ion. The latter correlation
is consistent with the earlier proposal that rate-determining steps involve OH
desorption.
Thin films of ABO2 metal oxides were prepared by RF-sputtering the parent
alloys under various O2 partial pressures, followed by annealing in air [122]. Metal A
was Pt or Pd, and B was Co, Rh, or Cr. The oxides exhibited the delafossite structure,
i.e., for PtCoO2, alternating layers of Pt linearly coordinated with two oxygen atoms
and Co octahedrally coordinated with oxygen. Oxygen evolution activity depended
strongly on the transition metal, in the order Co > Rh > Cr, but was nearly
independent of the noble metal. It was suggested that this was due to coverage of the
noble metal with a poorly conducting oxide. The Co containing oxides produced
oxygen evolution potentials more than 100 mV lower than a Pt anode in 1M NaOH at
23 oC.
In electrolysis at 120 oC NiCoO2 anodes prepared by thermal decomposition
showed oxygen evolution potentials similar to those of NiCo2O4 and La0.5Sr0.5CoO3
anodes at a current density of 1000 mA/cm2 [113,114]. At lower current densities
(100-200 mA/cm2), NiCoO2 was superior to the other anode materials.
It has been known for some time that nickel-iron alloys show relatively low oxygen
evolution overpotentials [123,124]. However, the variation of ηO2 with alloy
composition was not clearly defined, because results vary depending on alloy history
and pre-treatment.
Later Ni-Fe alloy anodes have been re-examined. Gras and Pernot [125]
studied the anodic behaviour of alloys containing 9, 28, and 50 w/o nickel, in
comparison to two forms of nickel. They found that the 9% Ni alloy was unstable, in
agreement with earlier results [124], while the catalytic activity of the 28% nickel
alloy was lower than that of pure nickel. The corrosion- resistant 50% nickel alloy
was about 70 mV more efficient than nickel after 100- 200h of electrolysis, in the
current-density range 100-1000 mA/cm2. Anode composition vs. depth profiling, by
secondary ion mass spectrometry, showed that the surface was enriched with
potassium and oxygen after electrolysis. It was proposed that in situ formation of a
NixFeyOzKt complex oxide increased the catalytic activity of the anode.
59
In another study [126], high surface-area nickel-iron alloy anodes were
prepared with a 37% nickel powder prepared by co-decomposition of nickel and iron
carbonyl. The powder was applied to anode substrates as a sintered, porous coating. In
30% KOH electrolyte at current densities up to mA/cm2, the alloy anodes evolved
oxygen as efficiently as similar anodes prepared with pure nickel powder.
Non-noble metal catalysts were made by plasma spraying coatings of >90 w/o
Ni and/or Co plus <10 w/o stainless steel onto preheated iron or nickel-coated iron
substrates [127]. Anode overpotentials were 60-80 mV lower than on smooth nickel,
at 1000 mA/cm2, in 29% KOH at 70 oC. At current densities <500 mA/cm2, there was
no variation of anode potential vs. time for 1000h of operation. At 1000 mA/cm2, the
anode potential increased from 1.54 to 1.60 V/NHE during the first 300h of
electrolysis, and then remained constant up to 1000h.
Ni-W alloys and the intermetallic compounds Ni3Ti, NiTi, and NiTi2 were
investigated by Lu and Srinivasan [128]. Titanium was alloyed with nickel to increase
the number of d-band vacancies and, thus, electrocatalytic activity [129]. However, as
discussed in the following section, such a correlation was not found at the oxidecovered electrodes. Best results were obtained with Ni3Ti, for which ηO2 was 20 mV
lower than on nickel, in 30% KOH electrolyte at 80 oC.
While the last-cited study found no correlation of oxygen evolution activity to
electronic properties, Osaka et al. [130] did find a correlation to magnetic properties,
which strongly connected, in turn, to electronic properties, for cobalt borides and
composite cobalt borides. For cobalt borides, the O2 evolution activity depended
significantly on the Co/B ratio, as well as the sintering temperature used to make the
catalyst. Optimum preparations produced higher electrocatalytic activity than nickel
or cobalt. These best conditions included a 3:1 ratio of cobalt to boron, and a sintering
temperature of 400o -500 oC at which the saturation magnetization showed a
maximum. Among the composite cobalt borides, cobalt iron boride (Co:Fe:B = 1:2:1)
sintered at 50 oC had the greatest catalytic activity. At 100 mA/cm2 in 6M KOH at 25
o
C cobalt iron boride produced a potential of about 0.65V vs. Hg/HgO. It was found
that those catalysts which formed thicker and more-stable oxide films in higher
oxidation states had higher activities for oxygen evolution.
The latter work established that nickel oxyhydroxide is the preferred
electrocatalytic species for oxygen evolution. Nickel oxyhydroxide forms when nickel
hydroxide is electrochemically oxidized to the +3 valence state. Appleby and coworkers demonstrated that oxygen evolution overpotentials were considerably lower
on porous sintered nickel when it was impregnated with nickel hydroxide80. It is likely
that nickel hydroxide impregnation produces more of the preferred electrocatalyst
nickel oxyhydroxide than is generated anodically on a metallic nickel surface.
Available experimental evidence strongly indicates that individual precious metals
and their oxides have little or no superiority to nickel in their electrocatalytic activity
[76,80,85,131]. The exception is ruthenium oxide, which is a highly efficient oxygen
evolution electrocatalyst [95] but is unstable in alkaline electrolyte.
Oxygen evolution on smooth platinum and iridium electrodes was examined
by Appleby et at. [80] and compared to O2 evolution on nickel. In 25% KOH
electrolyte, Ir was markedly superior to Pt in the temperature range 25o- 90 oC. In
34% KOH, the relative superiority of Ir was somewhat less. Tafel slopes were about
2RT/3F, with evidence of a change to a higher slope at current densities above 10
mA/cm2. However, neither Ir nor Pt was as effective an electrocatalyst as pure nickel.
Another study [76], in which Ni and Pt screen anodes were compared, produced the
same result.
60
Pt, Pd, Rh, and Ni coatings were electrodeposited onto foamed Ni anode
supports and tested at a current density of 200 mA/cm2, in 30% KOH electrolyte at 90
o
C [85]. The rhodium coating performed best, providing a cell voltage of 1.72V,
followed by Pd (1.74V), Ni (1.75V), and Pt (1.80V).
Ir and Ru alloys with 25, 50, and 75 a/o nickel were investigated as oxygen
evolution anodes in 30% KOH electrolyte at 80 oC [128]. The Ni-Ru alloys with 50
and 75 a/o Ru dissolved anodically. The maximum overpotential reduction was
produced by the 50Ni-50Ir alloy, which was 40 mV more efficient than pure nickel at
a current density of 20 mA/cm2. The electrokinetic parameters for the alloys were
similar to those for nickel. Correspondingly, cyclic voltammetry showed that the
surface of the alloy anodes was predominantly composed of nickel oxide species.
Because of coverage by oxide films, there was no dependence of electrocatalytic
activity on alloy electronic structure.
Outstanding oxygen evolution activity was obtained at pyrochlore structure
oxide anodes [132]. The catalysts are described by the general formula A2[B2-xAx]O7o
y, where A = Pb or Bi, B = Ru or Ir, O < x < 1, and O < y < 0.5. In 3M KOH at 75 C
4+
a typical
Pb2[Ru2-xPbx ]O6.5 catalyst evolved oxygen at an overpotential of
about 120 mV at a current density of 100 mA/cm2. In comparative potential sweep
experiments, pyrochlore anodes were >100 mV more efficient than Pt black, RuO2, or
NiCo2O4. A supported Pb2[Rul.49Pb00.514+]O6.5 s anode was life-tested at a current
density of 200 mA/cm2 for more than 1000h, in 3M KOH at 75 oC. After the first
200h, the rise in anode potential vs. time slowed noticeably, reaching a potential of
about 1.5 V/RHE, about 50 mV higher than the potential measured at the beginning of
the test. X-ray diffraction of the used anode did not reveal chemical degradation.
Attempts to maximize anode efficiency using high (62 mg/cm2) catalyst loading TFEbonded structures produced stability problems as the catalyst layer separated from the
support during 504h of operation.
A number of approaches have been taken to improve the oxygen evolution
anode. To date, results using metal oxide electrocatalysis have generally shown more
promise than those obtained with modifications of the nickel electrode structure.
Catalysts such as NiCo2O4, Li-doped Co3O4, and the La-Sr-Co oxides have been
tested extensively. Further research of this type may lead to important catalyst
modifications or the development of entirely new catalyst systems.
With the exception of the relatively inexpensive ruthenium, as used in the
pyrochlore oxides, it is unlikely that practical oxygen evolution anodes will be based
on precious metals. Overpotential reductions obtained with Pt, Ir, or Pd, for example,
has been modest. In addition, the promising results obtained with far cheaper
materials mitigates on economic grounds against the use of precious metals.
Electrolyte
The electrolytes are defined as the group of electrical second-class conductors (ionic
conductors). They are mostly found in liquid form, but also as solids [133- 136].
A differentiation is made between strong and weak electrolytes, depending on
their ability to dissociate in liquids more or less completely into ions. Those belonging
to the first category, which at the same time also form the group of electrolytes with
high electrical conductivity, include the mineral acids (H2SO4, HCl etc.) which are
also strong in the chemical sense, together with the strong bases (NaOH, KOH, etc.),
which are so important for water electrolysis. The weak electrolytes have no
significance in the field of electrolysis because of their low electrical conductivity.
1 1
61
In the conventional methods of water electrolysis, dilute caustic potash solution
(KOH) and dilute caustic soda solution (NaOH) are primarily used as electrolytes at
operating temperatures between 60 ° and 80 °C. They have the advantage that steel
and nickel can be used as the cheapest electrode materials, and also as the
constructional materials for the cell including the cell container, without any
manifestation of corrosion.
In addition, for some of the modified water electrolysis cells at present in development, solid electrolytes of zirconium oxide with admixtures of yttrium oxide or
ytterbium oxide as well as various organic and inorganic synthetics are being
investigated for their suitability as electrolytes [134-136].
3
4.2 Polymer electrolyzers
4.2.1 The cell
Proton exchange membrane (PEM) water electrolysis technology is frequently
presented in the literature as a very interesting alternative to the more conventional
alkaline water electrolysis.
Proton exchange membrane (PEM) water electrolysis systems offers several
advantages over traditional technologies including greater energy efficiency, higher
production rates, and more compact design [137]. This method of hydrogen
production is envisioned in a future hydrogen society whereby hydrogen as the energy
carrier is incorporated in an idealized ‘‘energy cycle’’. In this cycle, electricity from
renewable energy sources is used to electrochemically split water into hydrogen and
oxygen. The only input to this cycle is the clean renewable energy and the only output
is electric power. A basic schematic of a PEM water electrolysis cell is shown in
Figure 4-6. The PEM water electrolysis cell consists primarily of a PEM on which the
anode and cathode are bonded. These electrodes are normally a composite of
electrocatalytic particles and electrolyte polymer.
Cells that use a solid-polymer electrolyte are usually constructed on the filter
press-type design. They do not require electrolyte circulation because the electrolyte
is immobilized in the form of an ion-exchange resin. The electrodes are either
embedded in the surface of the resin sheets or pressed closely against the two
opposing faces of the sheet of resin material. A ribbed or corrugated solid metal
separator plate is interposed between cells, providing electric continuity between one
cell and the next while separating the hydrogen from the oxygen in adjacent cells.
This type of cell is usually cooled by circulating water through the cavity between the
metal separator and the electrode plate. Hydrogen or oxygen evolved into this cavity
is swept out by the coolant stream and is separated from the water outside the cell.
The advantages of the solid-polymer-type cell are that
(a) the electrolyte membrane or diaphragm can be made very thin, allowing
high conductivity without risk of gas crossover, and
(b) the electrolyte is immobilized and cannot be leached out of the cell.
Also it is ecological cleanliness, considerably smaller mass–volume
characteristics and power costs and, that is very important, a high degree of gases
purity, an opportunity of compressed gases obtaining directly in the installation, the
increased level of safety [138].
62
Figure 4-5. Schematic of a PEM electrolyzer cell
The disadvantages of the solid-polymer-electrolyte (SPE) cell are that
(a) the electrolyte costs more than the conventional alkaline solutions and
(b) the electrolyte is corrosive and requires more expensive metal components
to be used in the cell. For these reasons, solid-polymer-electrolyte cells are usually
operated at somewhat higher current densities than cells that use a liquid alkaline
electrolyte.
Normally, different electrocatalysts are utilized for the anode (e.g. IrO2) and cathode
(e.g. Pt). When the electrode layers are bonded to membrane, it is known as the
membrane electrode assembly (MEA). The electrical contact and mechanical support
is established with porous backings like metallic meshes or sinters. In a PEM water
electrolyser hydrogen is produced by supplying water to the anode where it Equation
3). The protons are transported through the proton conductive membrane to the
cathode. The electrons exit the cell via the external circuit, which supplies the driving
force (i.e. cell potential) for the reaction. At the cathode the electrons and protons recombined to give hydrogen gas (Equation (16).
Anode : 2H2O →4H+ + O2 + 4e-
(16)
Cathode : 4H+ + 4e- → 2H2
(17)
Cell : 2H2O → 2H2 + O2
(18)
63
Figure 4-6. The schematic diagram of the heart of a PEM electrolyzer.
The heart of a PEM electrolyzer is shown in the schematic diagram above. Its
efficiency is a function primarily of membrane and electrocatalyst performance. This
becomes crucial under high-current operation, which is necessary for industrial-scale
application.
The membrane consists of a solid fluoropolymer which has been chemically
altered in part to contain sulphonic acid groups, SO3H, which easily release their
hydrogen as positively-charged atoms or protons [H+]:
SO3H -> SO3- + H+
(19)
These ionic or charged forms allow water to penetrate into the membrane
structure but not the product gases, molecular hydrogen [H2] and oxygen [O2]. The
resulting hydrated proton, H3O+, is free to move whereas the sulphonate ion [SO3-]
remains fixed to the polymer side-chain. Thus, when an electric field is applied across
the membrane the hydrated protons are attracted to the negatively-charged electrode,
known as the cathode. Since a moving charge is identical with electric current, the
membrane acts as a conductor of electricity. It is said to be a protonic conductor. A
typical membrane material is sold by Du Pont under the trade name Nafion®. It has
several advantages over conventional electrolyzers which normally use an aqueous
caustic solution for workable conductivity. Because Nafion® is a solid, its acidity is
self-contained and so chemical corrosion of the electrolyzer housing is much less
problematic. Because it is an excellent gas separator, allowing water to permeate
64
almost to the exclusion of H2 and O2, it can be made very thin; typically only 100
microns, or one tenth of a millimetre. This also improves its conductivity so that the
electrolyzer can operate efficiently even at high currents. It is said that the membrane
suffers less from internal voltage losses due to a high current passing through a
smaller resistance, as given by Ohms Law, viz., V = IR.
However, the membrane also has some disadvantages. Unlike conventional
polymers which are water-repellent, Nafion® is a very expensive material. It must also
be kept humidified constantly, otherwise its conductivity deteriorates. This last is
never a serious problem in an electrolyzer because of contact with hot water, but the
PEM fuel cell requires intensive water management for stable, continuous operation.
4.2.2 Stacks and systems
PEM-based electrolysis systems offer a number of attributes such as modular
aspect (there is minimal penalty on efficiency due to unit size), all solid state system
(no alkaline liquid electrolytes or its recycling involved and water and electricity are
the only inputs required), pure hydrogen and oxygen generation (due to physical
separation by solid electrolyte membrane), ability to produce hydrogen at a pressure
(electrochemical compression), small footprint due to high current densities
achievable (>1 A cm−2 as compared to ~ 0.4 A cm−2 by alkaline system) [139- 141].
1
Figure 4-7. PEM Electrolyzer Systems (20 W to 20 kW). Hydrogen production: 5
l/h to 5 Nm³/h. Oxygen production: 2.5 l/h to 2.5 Nm³/h
65
Figure 4-8. Appearance of electrolysis test using 10 cells layered stack. (Cell
electrode area size: 1,000cm2) Operating pattern: Electrolysis 2Hr/Pause 30min.
Operating conditions: Temp.353K/Press.0.7MPa Current density: 1A/cm2
4.2.3 State of art
PEM electrolysis technology due to its fast response time and start-up/shut-down
characteristics (hydrogen generation starts immediately at ambient conditions) and
ability to accept large variations in load is ideal for integration with intermittently
available sources of electricity (renewable and off-peak grid). In addition, due to the
similar aspects of the PEM electrolysis and fuel cell technologies, the impact on
development and system cost reduction can be enormous.
Presently, the use of PEM water electrolysis systems, with only a few
commercial systems available [14,142- 145]. The restricting aspects of these systems
are the high cost of the materials such as the electrolyte membrane and noble metalbased electrocatalysts, as well as the complex system components to ensure safe and
reliable operation.
Development of PEM electrolyzers (the principle scheme of electrolyzer cell
with PEM is shown on Proton exchange membrane (PEM) water electrolysis systems
offers several advantages over traditional technologies including greater energy
efficiency, higher production rates, and more compact design. This method of
hydrogen production is envisioned in a future hydrogen society whereby hydrogen as
the energy carrier is incorporated in an idealized ‘‘energy cycle’’. In this cycle,
electricity from renewable energy sources is used to electrochemically split water into
hydrogen and oxygen. The only input to this cycle is the clean renewable energy and
the only output is electric power. A basic schematic of a PEM water electrolysis cell
is shown in Figures 6 is historically connected to development of perfluorinated ionexchange membranes Nafion from DuPont firm. The first PEM electrolyzers have
been created in 1966 by company General Electric146. Such electrolyzers were
1 1
66
developed for the special purposes (spacecrafts, submarines, etc.), and also for needs
of the civil industry.
Nafion® (E. I. du Pont de Nemours and Company), which is a
perfluorosulphonic acid membrane composed of polytetrafluoroethylene (PTFE)
backbone and pendant side chains terminating with -SO3-, has excellent chemical
stability and high protonic conductivity. Water electrolysis using the membrane as a
solid polymer electrolyte is a very promising method for large-scale hydrogen
production [147- 150]. In the method, electrocatalysts are tightly bonded on both sides
of the SPE membrane.
Now scopes of PEM electrolyzers, besides fuel cells, are analytical instrument
making, systems of correction of a water-chemical mode of nuclear reactors,
hydrogen welding, metallurgy of especially pure metals and alloys, manufacture of
pure substances for electronic industry, analytical chemistry (the equipment for a gas
chromatography, maintenance with hydrogen of laboratories), etc.
Research and development of electrolyzers with PEM were provided in Norway
(Norwegian University of Science and Technology [151,152]), in Spain (David
Systems and Technology) and also in Japan (Matsushita Electric Works, Ltd. [153]
and Fuji Electric Co., Ltd. [154- 156]). So, for instance, the certain progress has been
achieved within the framework of Japanese program WE-NET [155,156] (an element
with the area 2500 cm2, operating voltage (U) 1.556V at 80 oC and current density on
a visible surface (i) 1 A/cm2 with efficiency of energy transformation 95.1 % that is
explained by proximity to equilibrium potential 1.48 V). It allows speaking about
increase in specific productivity of PEM electrolyzers.
A lot of R&D was done in the field of PEM electrolyzers [146,151- 167], but
high price has limited their mass production. High cost of membrane (about 200$ per
1 m3/ h of hydrogen at i = 1 A/cm2), the electrocatalyst with noble metal (Pt, Ir, Ru),
high requirements to cleanliness of water and constructional materials (basically, Ti)
result in rather high cost of such type of electrolyzers. On the other hand, cost of
hydrogen produced by electrolysis usually approximately on 70 % consists of cost of
electricity. Therefore decrease of power consumptions of PEM electrolyzers
compensates relatively high capital expenses. The estimation, made for the same
service life (about 5 years), shows, that cost of the hydrogen made by PEM
electrolysis, even is less, than cost of the hydrogen made by an alkaline electrolysis, is
especial if to take into account the cost of buildings, auxiliaries, clearing up of
hydrogen and recycling of an alkaline solution.
Obviously, essential enough reduction of price is possible at increase in scales
of production and improvement of electrolyzers design. However, initially high cost
of such electrolyzers was an obstacle for development of their large scale
manufacture. It is necessary to note, that for manufacture of PEM electrolyzers the
same materials and technologies as for PEM fuel cells are used (for example, the
same membranes, catalysts on the basis of platinum metals, similar techniques of
catalyst synthesis and applying are used). From this point of view, there are real
preconditions for decrease of cost of PEM electrolyzers, related with the beginning of
large-scale PEM fuel cells production. So, for example, today wholesale party price of
Nafion membrane can already be reduced up to 200$/m2 and expected cost in the
nearest future is less than 200$/m2 (or less than 50$ for 1m3/h of produced hydrogen
at i = 1A/cm2).
Creation of different PEM electrolyzers and installations on their basis with
various productivity (from several millilitres up to tens cubic meters of hydrogen per
1 1
1
31 1 1 1 1 1 1 1 1 1 1
67
one hours, see Figures 9, 10, 11) was a result of R&D carried out in RRC “Kurchatov
Institute” during last 20 years.
Figure 4-9. Photograph of PEM electrolyzer with hydrogen productivity 20 l / h and
its components.
Figure 4-10. Photograph of PEM electrolyzer with hydrogen productivity 1.5m3 / h.
68
Figure 4-11. Photograph of PEM electrolyzer with hydrogen productivity 2m3 / h and
operating pressure up to 30 Bar.
For the present moment significant experience on designing and manufacturing
of laboratory and small-scale samples was saved up; there are design and
technological development for manufacturing of electrolyzers components
(membrane-electrode assemblies, current collectors, bipolar plates, sealing elements,
etc.).
Electrocatalysts
Since chemical (H2) energy is being created, a minimum energy must be input to
drive the process according to the laws of thermodynamics. In terms of electrical
energy, this corresponds to a voltage of 1.23V. In reality, the working voltage
necessary to sustain water electrolysis is always greater than this. The extra voltage,
generally known as the overvoltage, represents a waste of energy or loss of efficiency.
It has two main causes, one of which is the IR loss due to the finite electrical
resistance of the electrolyte, or membrane in this case (see above). The second is
kinetic in origin, i.e., to do with the overall speed of the process at the electrode
surface.
A solid catalyst (M) speeds up chemical reactions due to its surface action. As a
simple example, two H atoms held loosely on a surface are much more likely to
collide and make H2 gas than if they are dispersed in a liquid with billions of water
molecules in-between. This is a spatial or localized concentration effect. The case of
O2 evolution is much more complex. Two water molecules must be broken into their
constituent atoms; then the two O atoms must combine. The electrocatalyst at the
anode is a special catalyst which facilitates this process by withdrawing electrons
from the water such that the H atoms are ejected as protons, which enter the
membrane. Water is said to be activated by charge-transfer. The OH or O atoms are
69
very reactive in their free state. However, when fixed at the surface by chemical
bonds, they are much more stable. When more water encounters the surface, its
protons are ejected in turn and O atoms are accumulated. These are then able to
combine easily by surface diffusion just as described for hydrogen. It is said that the
surface provides a low-energy pathway or a new mechanism, which is intrinsically
much faster because the speed of the reaction is related exponentially to the energy
difference.
It is easy to visualize that if the cathode and anode surfaces, respectively, attract H or
O atoms too strongly, the surfaces will become completely covered with these
intermediates and the catalytic process stops. On the other hand, if protons or water
are not attracted strongly enough, the process never gets going. Only when there is a
moderate strength of binding of reactants and intermediates at the electrode surfaces
will the right balance is obtained. This is the key factor in determining if a solid
catalyst will work efficiently. It is also obvious that the larger is the catalyst surface
area available, the more H2 and O2 will be produced in a given time, i.e., a higher
current will flow in the electrolyzer.
Platinum is long known to be the best catalyst for water electrolysis due to its
moderate strength of adsorption of the intermediates of relevance. It has the lowest
over-voltage of all metals. Because of its cost, and the preferred operation of the
electrolyzer at high current, ingenious ways have been devised to deposit ultra-fine Pt
particles either on the electrode support plate, or directly onto the membrane, which is
then clamped for good electrical continuity. A current of 1-3 Amperes per square
centimetre can be obtained from as little as 3 milligrams of Pt spread over the same
area.
Electrocatalysts on the carriers, mixed oxide catalytic compositions were
developed, allowing to lower the loading of platinum metals without reduction of a
resource [168]. Thus it was shown, that activity of a composition with 40–50 at% of
RuO2 is comparable to activity of pure IrO2, and parameters of electrolysis with
RuO2(30%)-IrO2(32%)-SnO2(38%) as anode electrocatalyst at the loading of platinum
metals 0.8 mg / cm2 are practically similar to parameters of electrolysis with iridium
anode electrocatalyst with the loading 2.0–2.4 mg / cm2. Also rather good
performances have PEM electrolyzers with Pd catalysts (including Pd on carbon
carrier) on cathode. Performance data of electrolyzers are power consumption 3.9–4.1
kW-h / m3 of hydrogen at i = 1A/cm2, cell voltage U = 1.68–1.72V and t = 90 oC,
cleanliness of hydrogen more than 99.98 %; operating pressure up to 30 Bar; the
loading of noble metals 0.3–1.0 mg / cm2 on the cathode and 1.5–2.0 mg / cm2 on the
anode; operation time up to 10 000 h. Development of electrolyzers operating under
the pressure up to 150 Bar was provided.
As it was already marked, essential interest represents an opportunity of
carrying out PEM electrolysis of waters at the increased pressure. Researches have
shown [169], that for electrolysis at increased (up to 30 Bar) pressure the
improvement of volt–amperic characteristic was observed at operating current density
in comparison with electrolysis at atmospheric pressure. It is caused, first of all, by
reduction of an anode overvoltage. So, at i = 2 A / cm2 the voltage of electrolysis
decreases on 70–80 mV (Fig. 6). Besides this the increased pressure allows to carry
out electrolysis at temperature above 100 oC, thus reduction of power consumption
due to reduction of membrane resistance and decrease of an overvoltage take place. In
particular, at temperature 120 oC and pressure 25 Bar at current density 1 A / cm2 the
voltage on a cell was 1.65 V. The theoretical analysis allows supposing, that the
increase of pressure results in reduction of volume of gas bubble generated during
70
electrolysis that, in turn, promotes improvement of water transport, reduction of
ohmic losses in catalytic layer and improvement of electric contact between catalytic
layer and current collector.
Recently IrxSn1-xO2 [170-,,173], IrxRu0.5-xSn0.5O2 [174], and IrxRuyTazO2
powders have been investigated as oxygen evolution electrocatalysts. Normally in
these systems, the anode has the largest overpotential at typical operating current
densities (normally around 10 kAm-2) [175].
Noble metal oxides as electrocatalysts are well established in many industrial
electrochemical processes in the form of dimensionally stable anodes (DSA) as
developed by Beer [176]. Ruthenium is known as the most active oxide for anodic
oxygen evolution [177], however it suffers from instability and therefore should be
stabilized with another oxide such as IrO2 178 or SnO2 [179]. Tantalum is a wellknown addition to DSA electrodes, and Ir–Ta oxides have been suggested as the most
efficient electrocatalysts for oxygen evolution in acidic electrolytes due to the high
activity and corrosion stability [180]. Although very comprehensive studies of the
structural and electrochemical properties of DSA oxides have been carried out, little is
known regarding the structure and electrocatalytic properties of oxide powders. This
is mainly due to the prevalence of DSA electrodes in industrial processes compared
with the same oxide compositions in particle form. In addition, it is well known that a
wide range of preparation conditions affect the formation and properties of DSA type
oxide layers [181] as will the presence of Ti originating from the substrate [182], and
therefore analysis of powder-based oxides gives an insight to the specific nature of
noble metal oxides as electrocatalysts while providing a possible use in PEM water
electrolysis applications. By using a multi-disciplined approach to develop and
characterize electrocatalysts, Marshall [183] achieved significant improvements to the
performance and efficiency of a PEM water electrolysis cell. The best result was
obtained with an Ir0.6Ru0.4O2 anode and 20 wt % Pt/C cathode, with a cell voltage of
1.567 V at 1 Acm-2 and 80 oC when using Nafion 115 as the electrolyte membrane.
4.3 High temperature alkaline electrolyzers
In water electrolysis, water is split into hydrogen and oxygen, according to equation
(20), by an electric current which is passed between two electrodes submerged in an
electrolyte.
H2O + electric energy + heat → H2 +
1
2
O2
(20)
To split water, the electric voltage applied over the two electrodes must exceed a
minimum value: the so-called decomposition or reversible voltage. The reversible
voltage is determined by Gibbs free energy of water splitting and is thereby a function
of both pressure and temperature. At 25 ºC and 1 atm of the reactants the reversible
voltage is 1.23 V (see below).
Water is a very poor ionic conductor, and ions must be added in order to form
a conductive electrolyte so the reaction can proceed without too high resistance. Both
alkaline and acidic solutions can be used. In the following only alkaline electrolysis
will be described, where mainly potassium and sodium hydroxide solutions are used.
Of potassium and sodium hydroxide, potassium hydroxide possesses the lowest
resistance in the electrolyte and the lowest overpotential is archived. Therefore, in
most commercial water electrolysers, potassium hydroxide solutions (25-30 wt %) are
71
used. In the following, alkaline electrolysis based on a potassium hydroxide
electrolyte will be described. In alkaline electrolytes two basic reactions occurs at the
electrodes (equation (21) and (22)).
Cathode:
2 H2O + 2 e- → H2 + 2 OH-
(21)
Anode:
2 OH- →
(22)
1
2
O2 + H2O + 2e-
H2O → H2 +
Overall reaction:
1
2
O2
(23)
Most commercial potassium hydroxide water electrolysers use nickel electrodes and
are operated at 70-80 ºC. Increasing the operating temperature influences the
thermodynamic of the system, as will be described below. Only limited information
on electrolysers operated at elevated temperatures (above 150 ºC) is available. A few
authors have dealt with the theoretical thermodynamic aspects regarding electrolysis
at high temperature [184- 187] .
1 1
4.3.1 Thermodynamic of water electrolysis
In examining the thermodynamics of the electrolytic process, the electrolysis cell is
assumed to be ideal, consisting of a reversible hydrogen electrode and a reversible
oxygen electrode immersed in a solution of potassium hydroxide at a total pressure of
P and the temperature T. The hydrogen and oxygen in contact with the electrodes are
assumed to be wetted, containing water vapour in equilibrium with the water in the
electrolyte. Furthermore, the gasses are assumed to be ideal gasses.
The endothermic heat consumed by the splitting of water is related to the
enthalpy of formation of water ΔHf, which is described by Equation (24):
ΔHf = ΔG + T ⋅ ΔS
(24)
The total energy (ΔHf) needed for water splitting consists of a minimum fraction of
electrical energy (the Gibbs free energy ΔGf) and an entropy fraction ( T ⋅ ΔS f) as
shown in equation 5. Reaction 1 is endothermic, and the electrolyser will cool unless
heat ( T ⋅ ΔS f) is supplied.
4.3.2 Gibbs energy
The minimum thermodynamic voltage required for splitting of water is defined as the
minimum reversible voltage (Erev), which is related to the Gibbs energy of formation
of water. At standard conditions (25 ºC and 1 atm.) the minimum reversible voltage
is:
θ
Erev
=
1
n⋅F
ΔGθf, H2O
(25)
The standard Gibbs free energy of formation of water is: ΔGθf, H O = -237.13
2
KJ
mol
, n is
equal to two for splitting of water and F is Faradays constant (96485 As/mol). The
minimum required voltage at standard conditions for electrolysis of pure water (i.e. no
alkaline electrolyte) is thereby 1.23 V at 25ºC and 1 atm partial pressure for both
72
oxygen and hydrogen. The reversible voltage at any pressure and temperature can be
expressed by Nernst equation, which provides a relationship between the ideal
θ
standard potential ( Erev
) for the cell reaction and the reversible potential ( Erev ) at
other partial pressures of reactants and products:
Erev =
1
n⋅F
ΔGθf, H2O -
where Q is the reaction activity ( Q= P
1
n⋅F
PH2O
H2 ⋅PO2
½
R ⋅ T ⋅ lnQ
(26)
for reaction 4), which at standard state is
one for all reactants and products, thus Q = 1 and lnQ = 0 at standard state giving
equation 6.
Since the Gibbs free energy is a function of both temperature and pressure the
minimum voltage required varies with temperature and pressure as can be seen in
Figure 4-12, where Erev is shown as a function of temperature, pressure and
electrolyte concentration.
Increasing temperature reduces the reversible potential substantially. Thus it
can be advantageous to operate the electrolysis at elevated temperatures. On the other
hand, with increasing pressure the reversible potential increases corresponding to the
increase in free energy of the product gasses, as can be seen in Figure 4-12 and Figure
4-13. This increase is equal to the energy cost of compression at 100% efficiency.
Increasing the pressure can be beneficially regarding reduction of the relative volume
of the gas bubbles, which in turn increases the conductivity of the electrolyte (see
below).
Minimum thermodynamic voltage (U
, V)
1.4
25 ºC
50 ºC
100 ºC
150 ºC
200 ºC
250 ºC
1.3
1.2
1.1
1
0.9
0.8
0
5
10
15
20
25
30
Pressure (atm)
Figure 4-12. The minimum reversible voltage (Erev) required for the electrochemical
splitting of water as a function of temperature and pressure for pure water.
73
4.3.3 Enthalpy
For water electrolysis, the enthalpy voltage is defined based on the enthalpy of
formation of liquid water at the operation temperature and pressure. The minimum
thermodynamic voltage (also called the enthalpy voltage), corresponds to the change
in enthalpy during the electrolysis process, at the temperature of the electrolyte. The
molar enthalpy thereby corresponds to the enthalpy voltage or the more popular name
the thermo-neutral voltage (Etn).
Etn = - n1⋅F ⋅ ΔH
(27)
where ΔH as the molar enthalpy of formation.
As Erev, Etn decreases with increasing electrolyte temperature; at 25 ºC, Etn is
1.48 V, and decrease to 1.43 V at 300 ºC. The effect of pressure on both oxygen and
hydrogen is zero since they are assumed to be ideal gasses. In practice the enthalpy
voltage for water has a slight dependence on the operating pressure due to the nonideal behaviour of the gasses.
The thermo-neutral voltage is the voltage at which a perfectly insulated
electrolyser would operate, if there were no net inflow or outflow of heat. The value
of the thermo-neutral voltage is thereby a thermodynamic quantity, as a function of
the operating conditions of electrolyte temperature, electrolyte concentration, and
total pressure.
1.6
1.5
1.4
Thermoneutral voltage (Etn)
Entropy energy
Voltage
1.3
1.2
1.1
Gibbs' energy
1
Reversible voltage (Erev )
0.9
0.8
0
50
100
150
200
250
300
Temperature [ºC]
Figure 4-13. Reversible and thermoneutral voltage and as a function of temperature.
As mentioned above, conventional electrolysers are operated at 70-80 ºC and a
voltage of 1.7 – 1.9 V, thereby providing more electrical energy than thermodynamic
necessary (above the thermoneutral voltage). The excess energy must be removed
from the system by cooling water. The electrolysis voltage is the sum of the reversible
voltage and a contribution, which accounts for losses due to the overpotentials in both
electrodes and ohmic losses in the electrolyte as shown in Figure 4-14.
74
Figure 4-14. Prediction of voltage-current density performance for a 28w% KOH
electrolyser operated at 70 ºC [188].
The ohmic loss in the electrolyte significantly contributes to the increased cell voltage
necessary for the electrolysis. The conductivity increases up to 150 ºC, i.e. the ohmic
losses in the electrolyte decreases with increasing temperature. Figure 4-15 shows the
temperature dependence on the electrical conductivity for KOH and NaOH
electrolytes.
Figure 4-15.Temperature dependence of the electrical conductivity of aqueous
solutions of KOH and NaOH [189].
Because of the increased conductivity with temperature, a rise in the operation
temperature would allow for significant energy savings. For NaOH, the conductivity
passes through a maximum at around 150 ºC; an operating temperature above this
limit would no longer be beneficial regarding the electrical conductivity of the
electrolyte solution. On the other hand, for 50 %wt KOH, the highest conductivity is
75
found at temperatures above 200 ºC. Because of the high vapour pressure at elevated
temperature it is necessary to operate the electrolyser at high pressure. With
increasing pressure the relative volume of the formed hydrogen or oxygen gas bubbles
is lowered. The bubbles cause ohmic losses, since the electrical conductivity of gasses
is zero. Thus, the ohmic losses caused by bubbles are minimized at high pressure. On
the other hand, with increasing pressure, gas accumulation within the porous electrode
materials can occur. The consequence of gas bubbles within the porous electrode is
that only a small fraction of the interfacial surface is efficiently used [190].
Consequently highly porous electrode materials are of great interest for electrolysers
operated at high pressure. Increasing pressure is generally advantageous [191], where
the increased minimum reversible voltage is cancelled by the lower ohmic loss due to
the decreased relative volume of the gas bubbles, and faster kinetics of the electrode
processes, i.e. lower overpotential. As described above, increasing the operating
temperature would obviously increase the efficiency of water electrolysis. However,
information on electrochemical studies of hydrogen and oxygen evolution reactions in
alkali solutions at elevated temperatures is scarce. As described above, a few authors
have dealt with the theoretical aspects regarding high temperature electrolysis [184187]. To the best of our knowledge only very limited experimental data for alkaline
electrolysis at temperatures at 150 ºC and above have been reported [192- 194].
Nevertheless it was proven that the efficiency of water electrolysis for the production
of hydrogen over polished nickel electrodes increases significantly by high
temperature operation as shown in Figure 4-16 [193].
3
1
Figure 4-16. Potential vs. current density for hydrogen and oxygen evolution on
polished nickel electrodes in 50 wt% KOH solutions at temperatures from 80 ºC to
264 ºC [193].
76
From Figure 4-16 it is clear that an increase in temperature has a more pronounced
effect on the oxygen evolution than the hydrogen evolution reaction. Increasing the
temperature significantly shifts the oxygen evolution reaction to lower potentials,
which was explained by the slow kinetics of the oxygen electrode reaction [193]. For
the hydrogen evolution reaction, only small shifts in potential were observed at
temperatures above 150 ºC at low current densities. Nevertheless, substantial
overpotentials are found for the hydrogen evolution reaction on nickel electrodes in
alkaline solutions at lower temperatures [193]. For the oxygen evolution reaction the
potential at a current density of 0.3 A/cm2 decreases from 1.73 volts to 1.26 volts by
increasing the temperature from 80 ºC to 264 ºC (1.59 volts at 150ºC and 1.41 volts at
108 ºC), whereas for the hydrogen evolution reaction, a slightly smaller decrease was
observed from -0.44 volts at 80 ºC to -0.18 volts at 264 ºC (-0.34 volts at 150ºC and 0.24 volts at 108 ºC) over the polished nickel electrodes, as shown in Figure 4-16.
The dual region of the Tafel slope was explained by a change in mechanism due to the
magnetic properties of nickel, although at present the exact mechanism is not
understood. Regardless of the interpretation of the Tafel slopes, it is evident that
increase in temperature significantly improves the kinetics of both oxygen and
hydrogen evolution, and electrolysis at elevated temperatures would be beneficially.
As can be seen from Fig. 3, also overpotential related to the ohmic resistance
of the electrode materials contribute significantly to the electrolysis voltage. The
known materials for alkaline electrolysis are very limited because of corrosion, and
they become increasingly limited at elevated temperatures.
4.3.4 Materials
Both platinum and palladium shows lower overpotential for oxygen evolution than
nickel, but because of economical reasons, nickel is almost exclusively used in
conventional electrolysers [195,196]. To decrease the overpotential, several alloys
(Ti-Ni [197]and Ni-Ir, Ni-Ru, Ni-Mo [198]) were tested and showed slightly lower
overpotential. For pure nickel electrodes, mainly Raney-nickel, i.e. leached nickel is
used, because of the high active surface area and porosity. Other highly porous
materials such as Pt/C was suggested as alternatives to the nickel electrodes [199201] .Nickel materials have also shown relative low overpotential for hydrogen
evolution.
Of more advanced materials, spinel and perovskites (Co3O4, NiCo2O4, LaNiO3
and La0.5Sr0.5CoO3) were shown to perform better as oxygen electrode (anode) than
pure nickel, although Raney-nickel was shown to perform with lowest overpotential
in one study [202]. Contrary, other studies showed that the mixed metal oxides, and
especially spinel and perovskites (Co3O4, NiCo2O4, LaNiO3 and SrCoO3) were the
preferred anode materials [203,204].
Most electrode materials have been tested at temperatures up to 160 ºC with a
slight increased thermal degradation at high temperature [195-197,203,204].
Conventional nickel electrodes was shown to withstand temperatures up to 200 ºC, as
they were used for the Apollo Fuel Cell System [205].
The separators/diaphragms with the electrolyte serve as an ionic conducting
material as well as separating the product gasses. Separation of the product gasses
becomes increasingly important at high pressure where oxygen become highly soluble
in the electrolyte (KOH). For low temperature electrolysers, the separator/diaphragm
can be nickel oxide, asbestos or polymer. Both the polymers and asbestos become
instable at temperatures above 120 ºC [189]. Therefore at high temperature
2
3
77
electrolysis new separator/diaphragm materials should be developed. Oxide-ceramic
diaphragms such as ceramics of titanates (BaTiO3 and CaTiO3) or even NiO
[189,206,207] can be suitable substitutes for the polymeric or asbestos
separators/diaphragms at high temperatures.
The construction material for the electrolyser has to withstand the enhanced
general corrosion as well as stress corrosion at high temperature operation. Titanate
based ceramics or titanate coated Ni-alloys can probably be used for construction of
commercial alkaline electrolysers operating above 200 °C, but it remains to be proven
over periods of several years.
4.3.5 Conclusion
Increasing the operation temperature for alkaline water electrolysis significantly
increase the efficiency. A possible obstacle for operating at elevated temperature is
the lower stability of the materials. The biggest problem in designing and developing
a large scale water electrolysis plant to operate at elevated temperatures will be the
development of suitable substitute materials for the separator material. Thus the
question to be answered considering electrolysis at elevated temperature is whether
the efficiency benefit made by the high temperature operation is sufficient to
compensate for increased research and development. At present suitable materials
have been identified which are not more expensive than existing separator materials.
4.4 Solid Oxide Electrolyser Cells
4.4.1 Introduction
Electrolysis is a 200-year-old method for hydrogen production, and still electrolysis is
presently, and for the foreseeable future, the only method of practical importance for
hydrogen production by splitting of water. The chlorine-alkaline electrolysis, which is
worldwide the largest source of electrolytic produced hydrogen, has been in
commercial use for about 100 years. In this process hydrogen is regarded as a byproduct and chlorine is the main product. Only a vanishing small portion (of the order
of 0.1 %) of the world production of hydrogen is produced directly by water
electrolysis. Even this small quantity has been declining during the recent years since
the electrolytic production of hydrogen for fertilizer manufacture is not competitive
with production from natural gas. [208].
Recently, a strongly increased interest in hydrogen and CO2 neutral energy
production has aroused, [209,210] and very enthusiastic but not necessarily realistic
visions have been published. [211]. The hydrogen economy vision has been rejected
by Bossel et al. [212,213] and the rejection is convincing. From Bossel’s reports one
may get the impression that it does not make sense at all to use electrical energy for
electrolysis for hydrogen production. Bossel’s main argument is that there is a big loss
of the order of 75 - 80 % in converting electricity into hydrogen and back to
electricity again, whereas there is only about a 10 % loss in transporting electricity by
the grid. This argument is also correct in its essence even though the exact loss in
converting electricity into hydrogen may be lower for future electrolyser generations
than the usually assumed 30 %. Another feature pointed out by Bossel is that
molecular hydrogen is troublesome and expensive to handle.
78
We think that in spite of these arguments there might be a future market for
efficient electrolysers, which can split not only water but also carbon dioxide and
produce synthesis gas, a mixture of hydrogen and carbon monoxide. All types of
hydrocarbon based fuel may be produced from H2 and CO. Especially it is
inexpensive to produce the simplest synthetic fuels, namely methane, CH4, the main
constituent in natural gas, and methanol, CH3OH. Also synthetic petrol and diesel
may be (and have been in Germany during Second World War) produced from
synthesis gas in large quantities using the Fischer-Tropsch method. [214] Today new
large plants for manufacturing of synthetic diesel from synthesis gas using FischerTropsch are being built due to the high oil price. The synthesis gas is made by steam
reforming of cheap natural gas, which is available in certain places, e.g. the Near East.
Synthesis gas may also be produced from coal. It is generally believed that production
of synthetic fuels will be profitable, if the price of crude oil will be stable above 50
US$ per barrel.
The question “is there a possible market for efficient electrolysis?” can only be
answered by economic assessments. It is clear that the market will only be there if one
of two conditions will be fulfilled: 1) the price per unit energy of fossil fuel is
significant higher than the price for alternative energy like renewable energy (wind,
solar, hydropower) or nuclear energy; or 2) fossil fuel consumption is restricted by
political means. As a first step in answering if is possible to fulfil condition 1), the
hydrogen and CO production price will be discussed below. The discussion assumes a
CO2 price; however the price of CO2 of a reasonable purity is a complicated story and
will not be discussed in detail here. Instead a discussion on possible methods to
extract CO2 from the air is given after the discussion on production prices. However,
first a brief review of reversible solid oxide cells (SOC) is given followed by a status
of the state of the art of the SOC technology.
4.4.2 SOEC History and Background
The reversibility of solid oxide fuel cells (SOFC), i.e. that they can also work in the
solid oxide electrolyser cell (SOEC) mode, was proven already 25 years ago by A.O.
Isenberg. [215] Electrolysis of both water (steam) and CO2 was demonstrated.
Recently, the increasing interest in hydrogen production has created further interest in
the solid oxide cells (SOC) as electrolyser. During the time since the work of Isenberg
the power density of the SOC has increased significantly, or in other words, the area
specific internal cell resistance has been decreased substantially at least for fresh cells.
In this section we give a short introduction to the SOEC technology and present a
review of the international state of the art of SOCs.
High temperature electrolysers were under development during the 1980'es.
One advantage of the high temperature is that a part of the energy required for water
splitting is obtained in the form of high temperature heat, and thus the electrolysis is
performed with lower electricity consumption. The discussions focused on the use of
heat from solar concentrators or waste heat from power stations for this purpose
[216]. Due to a low energy price this development was stopped around 1990. The
high temperature solid oxide electrolyser cell (SOEC) has the advantage that it can
also split CO2 into CO and O2. Further, the high temperature is speeding up the
reaction kinetics, which in turn decreases the internal cell resistance and, thereby,
increases the energy efficiency. These features open up new potential possibilities for
a broader application of renewable or nuclear energy in the future, if fossil fuels
79
become scarce, and therefore several R&D-projects on SOEC are now being started
again both in Europe, USA and Japan. [210,217- 219].
A system consisting of a heat exchanger and a reversible SOC system has
clear advantages compared to low temperature electrolysis. Because both H2O and
CO2 electrolysis are increasingly endothermic with temperature, electricity demand
can be significantly reduced, if the formation of hydrogen is taking place at high
temperatures (600-1000 °C) as discussed in the section about fundamentals and to be
discussed further below. The electric energy need is reduced because the unavoidable
Joule heat of an electrolysis cell is utilized in the electrolysis process at high
temperature. If heat is available from sources such as heat of geothermal (e.g. on
Island), solar or nuclear origin, this will further reduce the electric energy demand for
hydrogen production by steam electrolysis. Even where such high temperature heat is
not available SOEC may be of interest. All heat sources with temperatures above 100
°C (the boiling point of water) are extremely beneficial since electric energy for steam
rising will be saved.
The Faradaic efficiency of SOEC has been shown to be 100 % over a period
of 1000 h [216], i.e. there are no parasitic reactions. This taken together with the
endothermic nature of the electrolysis process means that the H2 or CO efficiency,
defined as the total chemical energy (enthalpy of reaction, ΔH) in the H2 or CO
divided by the electric energy consumed, will be 100 % minus the heat loss from the
electrolyser to the surroundings. Thus, for well-insulated SOEC stacks in systems in
the range of 1 MW or above (ca. 1 m3 stack volume) the thermal loss can most
probably be well below 10 %. Also, some electric energy will be consumed in the
system for inverters and pumps, but again for reasonably sized systems this may be
few % only. This means that the SOEC technique has a potential efficiency of ca. 90
% for a system.
2
Figure 4-17. Sketch of an SOEC system for synthetic fuel production by electrolysis
of steam and CO2. CO2 and H2O are fed through the heat exchanger to the cell. Here it
is split into H2 and CO (syngas) and O2. On the way out, the synthesis gas is catalysed
into synthetic fuel using a catalyst.
As will be discussed below SOEC can split carbon dioxide into carbon monoxide and
oxygen, and the CO2 splitting has endothermicity similar to that of water splitting.
This means that electrolysis of a mixture of steam and carbon dioxide results in a
mixture of hydrogen and carbon monoxide called synthesis gas or short: syngas. By
catalytic reactions a number of other energy carriers may be produced from syngas.
The two simplest are methanol and methane. The preferred catalyst for CH4 formation
is Ni. Since the negative electrode of a SOC is partly made of Ni it is in principle
80
possible to produce CH4 within the cell (at high pressure and low temperature)[12]. The
entropy change for CH4 production from CO2 and H2O is nearly zero. This means that
the overall efficiency for a conversion of electricity to CH4 and back again can be
very high, if the reaction kinetics are fast, since only small reaction entropy losses
occur. The catalytic reaction to form CH4 or CH3OH from syngas can also be done in
a heat exchanger after the cell as sketched in Figure 2. This means that the energy for
H2O vaporization can be produced within the system. A combination of the two ways
to produce CH4 may prove to be the best production method, since it seems to
optimize efficiency and production rate.
4.4.3 SOEC state of the art at Risø
The SOC of all types are basically reversible cells and can be operated as solid oxide
fuel cells (SOFC) for electricity production and as well as solid oxide electrolysis
cells (SOEC) for production of hydrogen and synthesis gas. Figure 3 presents the
kinetics for SOCs fabricated and tested at Risø in both fuel cell mode (SOFC) and
electrolyser mode (SOEC) at different temperature and steam partial pressure. [220] It
illustrates that the cell is genuinely reversible as the I-V curves go smoothly through
the zero-current-density-point.
Figure 4-18. Kinetics of a Risø SOC working as an electrolyser cell (negative current
densities, i) and as a fuel cell (positive current densities, i) at different temperatures
and steam or CO2 partial pressures in the inlet gas to the cell. [220]
At a cell voltage of 1.48 V the produced joule heat within the cell equals the
consumed heat in the steam generation plus the steam electrolysis process. 1.48 V is
therefore called thermoneutral potential (Etn). At Etn and 950 °C with 70% H2O + 30%
H2 in the inlet gas a current density of -3.6 A/cm2 was measured with 30% steam
utilization. To the authors best knowledge this is the highest current density reported
in the literature for SOEC operation. Also included in Figure 3 is an i-V curve at 950
°C with 70% CO2 + 30% CO in the inlet gas. At -1.5 A/cm2 the cell voltage was 1.29
V and the CO2 utilization was 21 %. The internal resistance (the slope of the i-V
curve) is almost as good in electrolyser mode as in fuel cell mode.
81
A low internal resistance of the cell both at start-up and during thousands of
hours of electrolysis operation is important if SOCs should become interesting from a
commercial point of view, because the hydrogen production price is dependent on the
resistance of the cell. So far only few results on durability of high performing SOECs
have been reported in the literature. Even though the operation of the SOCs is
reversible and have comparable initial performance in electrolysis and fuel cell mode,
the degree of degradation (or passivation) of the cells during long-term testing in fuel
cell and electrolysis operation mode can be dramatically different as seen by
comparison of test of the same types of cell in the two modes [221,222]. Whereas the
cell in fuel cell mode is reasonably stable over years with high current density 1 - 2
Acm-2, it is in the electrolyser mode only stable at relatively low current density at 0.5
Acm-2, and even this has not yet been demonstrated over years. Figure 4 shows that
the cell voltage first increases slightly and then is stable over 1500 h.
1.2
Cell Voltage (V)
1
0.8
0.6
0.4
0.2
0
0
200
400
600
800
1000
1200
1400
Time (h)
Figure 4-19: Cell voltage measured during 1500 hours of electrolysis testing.
Experimental conditions were kept constant at −0.5 A/cm2, 850°C, p(H2O) = 0.5 atm
and p(H2) = 0.5 atm to the hydrogen electrode and O2 was passed over the oxygen
electrode. The steam utilization was 28%. Time zero is the point of time where −0.5
A/cm2 is applied.
Strong indications have been found that the passivation in fuel cell mode is due to an
accumulation of impurities at the three phase boundaries. [222,223] The phenomenon
is under further study at Risø. We believe strongly that this passivation can be
handled by proper handling of the trace impurities in the hydrogen electrode.
However, the impurities seem to be silica originating from the glass sealing, and thus
the durability may be increased by improving the sealing.
The degradation problem at high current densities must be solved before an
acceptable SOEC lifetime can be achieved. This may take a considerable R&D effort
of many man-years over several calendar years.
82
4.4.4 International SOEC status
A number of SOEC topics are currently being discussed in the literature such as
diffusion limitations, [224,225] heat management, [226] performance and durability,
[223,227,228] impurity studies, [229]microstructure, ball milling and preparation,
[230,231]performance of LSM, LSF and LSCo anodes, [232]ceria based composites,
[219,233] application of proton conductors, [234]system studies, [235,236]and life
cycle assessments. [209]
Table 4-1 summarizes the literature results of area specific internal resistances
(ASR) of SOECs at similar conditions. The variation is large, but by examination of
the references it seems that there is some correlation to the year of publication. In
other words, the SOCs have been improved very significantly over time due to the
large international R&D efforts. There are good reasons to believe that these
improvements will continue to happen during the coming years. A discussion on the
H2 and CO production prices using SOEC technology is given in the next section.
Table 4-1. Some reported initial performances of electrolysis cells. Comparison of
ASRs obtained from i-V-curves. The ASRs are taken as the slopes in the linear
regions of the electrolysis i-V-curves presented in the references sited. For each
reference is given the ASR measured on full cells with experimental conditions best
matching the ones of Figure 4-18.
Ref.
[223]
[223]
[216]
[237]
[238]
[239]
[240]
[219]
[241]
T
[°C]
850
950
1000
908
1000
1000
850
900
850
p(H2O)
[atm]
0.50
0.50
0.67
0.67
0.91
0.50
0.50
0.50
0.11
p(H2)
[atm]
0.50
0.50
0.33
0.33
0.09
0.50
0.50
0.50
0.89
ASR
[Ωcm2]
0.27
0.15
1.17
2.7
2
0.7
0.45
1.8
0.35
Specifications
Ni/YSZ-YSZ-LSM planar
Ni/YSZ-YSZ-LSM planar
Ni/YSZ-YSZ-LSM tubular
Ni/YSZ-YSZ-LSM tubular
Ni/YSZ-YSZ-LSM
Ni/YSZ-YSZ-LSM
Ni/YSZ-ScSZ(175 μm)LSM
Ni/SDC-YSZ-LSC
Ni/YSZ-ScSZ(125 μm)-LSM
As it is well known from the SOFC literature, the properties and performance of cells
are very dependent on the exact details of the fabrication procedure. This gives rise to
a lot of apparent contradiction in the literature such as several claim of LSM being a
bad oxygen electrode in contrast to the Risø findings that LSM is an excellent and
very durable electrode.
In conclusion it may be stated that of the SOC tested in SOEC mode and
reported in the literature the Risø cell have the best performance.
83
4.5 Economic modelling of H2 and CO production using SOEC
As already mentioned the high temperature steam electrolysis using SOECs has
gained renewed interest during the last few years. [209,210,227,230,231,235,242246]
2 2 2
A simplified model system is used to make an economic modelling, which has mainly
two purposes: 1) to assess the commercial potential of the SOEC, i.e. is it at all
conceivable that SOEC systems may become commercial? and 2) to make a
sensitivity analysis. The direction for SOEC optimization is trivial from a technical
point (the internal resistance should be as low as possible, the lifetime as long as
possible, the materials as cheap as possible etc). Thus, the only way of making a
reasonable priority of the R&D efforts is to find out which technical parameters are
influencing the economy the most.
At the present stage of R&D the actual economic figures are very uncertain
and should only be regarded as best estimates.
Some of the input is based on the results of a cell test at Risø. An SOEC of 8
2
cm Ni/YSZ (yttria stabilized zirconia) supported solid oxide cell with a 300 μm thick
support layer, a 10 μm thick Ni/YSZ electrode, a 10 μm thick YSZ electrolyte and a
20 μm thick LSM(strontium doped lanthanum manganate) /YSZ electrode was tested
at ambient pressure in a setup where the cell was sandwiched between two alumina
blocks. [247- 250]
The production cost below is given in equivalent crude oil prices using the
higher heating value (HHV) of the H2, the CO and the crude oil. For comparison it
can be mentioned that 100 US¢/kg H2 app. corresponds to 43 $/barrel crude oil using
the higher heating value (HHV). 10 US¢/kg CO corresponds to app. 60 $/barrel crude
oil, again using the HHV.
2 2
4.5.1.1 System description
The system used to estimate H2 production cost is shown in Figure 4-20. Water is fed
to the system where it is demineralised and subsequently evaporated by heat supplied
from an external heat reservoir. In order to keep the Ni in the Ni/YSZ electrode
reduced, a small part of the produced H2 is recycled. The inlet gas is further heated in
a heat exchanger by the hot outlet gas. Finally at the SOEC stack the water is
electrolyzed into H2. The same system sketch applies to the estimation of CO cost,
except no CO2 evaporation is involved.
Figure 4-20. Sketch of high temperature steam electrolysis using an SOEC stack. The
same sketch applies to CO2 electrolysis except no CO2 evaporation is involved.
84
If steam and CO2 electrolysis is combined, the produced synthesis gas can be
catalyzed into various types of synthetic fuel. The heat generated in the catalysis
reaction can be utilized for steam generation which means the heat reservoir becomes
more or less superfluous. Figure 4-21 shows a sketch of a steam and CO2 electrolysis
system combined with a synthetic fuel catalyst. The hydrocarbon fuel produced at the
catalyst may be condensed and the left-over synthesis gas may be recycled, and this
will keep the Ni in the Ni/YSZ-electrode reduced.
Figure 4-21. Sketch of SOEC steam and CO2 electrolysis combined with synthetic
fuel production. The heat generated in the catalysis reaction may be utilized for water
evaporation.
4.5.1.2 Thermodynamics and pressurization
The thermo-neutral voltage is defined as
ETn =
ΔH f
nF
(28)
where ΔH f is the formation enthalpy, n is the number of electrons involved in the
reaction and F is faradays constant. At ETn the heat consumption from the
endothermic electrolysis reaction equals the produced Joule heat within the SOEC i.e.
no surplus heat (waste heat) is produced and the stack needs no cooling.
For CO2 electrolysis, ETn = 1.46 V at 0.1 MPa, 950 °C. For steam electrolysis,
ETn = 1.29 V at 950 °C, 0.1Mpa. Hence, electrolysis of a steam/CO2 mixture can be
performed at thermo-neutral conditions at a cell voltage between 1.29 V and 1.46 V
depending on the steam electrolysis/CO2 electrolysis ratio.
As mentioned above, both the CO2 and the steam electrolysis reaction
becomes increasingly endothermic with temperature which is also seen in Figure 4-22
for steam electrolysis. This figure shows the energy demand for the steam electrolysis
reaction with 99% H2O + 1% H2 at the Ni/YSZ electrode and 100% O2 at the
LSM/YSZ electrode at 1 atm and 200 atm. Note how the evaporation temperature and
the electric energy demand, ( -ΔG f ) , increases with pressure. The data used to
construct the figure was acquired from NIST Chemistry Webbook. [251]
85
Figure 4-22. H2O splitting energies calculated at 1 and 200 atm. with data acquired
from Nist Chemistry webbook. [251].
4.5.2 The experimental results used as input
The top graph of Figure 4-23 show SOEC i-U curves obtained at various temperatures
and gas compositions. Details are given in Table 4-2. The bottom graph show the area
specific resistance corrected for gas conversion which is found from the i-U curves.
The CCASR calculation method is described in Appendix 1 in this chapter. The
CCASR is fairly independent of the current density. Hence, average values of the
CCASR are used to estimate the CCASR at any given temperature between 650 °C
and 950 °C. The average values are plotted in the inset vs. 1000/T and fitted with a
third order polynomial.
Table 4-2. i-U curve operating conditions.
i-U curve
A
B
C
*D
*E
Temperature
650 °C
50
10
25
0.991
750 °C
50
23
25
0.960
850 °C
50
50
25
0.929
950 °C
72
37
45
0.881
950 °C
70
#
22
33
0.868
1.53
0.53
0.23
0.16
0.23
1.53
0.53
0.22
0.15
0.22
H2O conc. [vol%]
Max H2O utilization [%]
Total gas flow rate [l/h]
OCV [V]
+
ASR [Ω cm2]
+
2
CCASR [Ω cm ]
#
*LSM/YSZ electrode gas was 20 l/h O2 instead of 140 l/h air.
#
Ni/YSZ electrode gas composition was CO2/CO instead of H2O/H2.
86
+
Area specific resistance and conversion corrected ASR is described in Appendix 1 in
this chapter.
In Table 4-2 is both shown the ASR and the CCASR values. Due to the limited
utilization ratios the difference between the ASR and the CCASR values are very
small. At higher utilization ratios the difference becomes more significant. For
instance, if the ASR is 0.2 Ωcm2, the inlet gas consists of 95% H2O + 5% H2 and the
outlet gas consists of 95% H2 + 5% H2O, the CCASR is only 0.13 Ωcm2. The CCASR
parameter is more adequate for modelling than ASR since it allows for a more precise
estimation of the H2 or CO production rate at utilizations differing from the
experimental ones.
Only one i-U curve was recorded with CO2/CO. The ratio between the
CCASR values of D and E in Table 4-2 is 1.5. Hence, as an estimation of the average
CCASR value of CO2 electrolysis is taken the third order polynomial for H2O
electrolysis multiplied by 1.5.
Figure 4-23. Top graph: i-U curves obtained at different temperatures. Details on gas
compositions are given in Figure 4-2. Bottom graph: Area specific resistance
corrected for gas conversion at the electrodes (CCASR) The inset show average
CCASR values vs. 1000/T. These values are fitted by a third order polynomial (black
line).
4.5.3 Economic input
Table 4-3 shows the assumptions for the H2 and CO cost estimations. Electricity from
nuclear power sources has been estimated to be 2.1-3.1 US ¢/kWh. [219] Non-firm
(or secondary) geothermal electricity prices for the power intense industry at Iceland
lay between 1-1.4 ISK(2002)/kWh from 1990 to 2002 with an average about 1.2
ISK(2002)/kWh. [252] This corresponds to 1.3 US¢/kWh with an average 2002
87
exchange rate between ISK(2002) and USD(2002) of 0.011. In Norway, between
1997 and 2003, the average cost price for the iron/steel and ferro alloy industry was
11 øre/kWh again corresponding to 1.3 US ¢/kWh. [253] This reflects the choice of
the electricity price.
Table 4-3. Cost estimation input parameters
Electricity price
Heat price
Investment cost
Demineralised Water cost
CO2 cost
Interest rate
Life time
Operating activity
Cell temperature
Heat reservoir temperature
Pressure
Cell voltage (H2O electrolysis)
Cell voltage (CO2 electrolysis)
Energy loss in heat exchanger
H2O or CO2 concentration in inlet gas
H2O or CO2 utilization
1.3US¢/kWh (3.6US$/GJ)
0.3US¢/kWh
4000 US$/m2 cell area
2.3 US$/m3
2.3 US$/ton
5%
10 years.
50%
850 °C
110 °C
0.1 MPa
1.29 V
1.47 V
5%
95%
95%
A 5kW plant based on SOFC technology is predicted to cost 350-550US$/kWe. [254]
Assuming a power output of 1W/cm2 this corresponds to an investment cost of 35005500US$/m2 cell area. As input for our estimations we choose 4000 US$/m2 cell area.
Demineralised water costs with systems such as the Recoflo system with a capacity of
45m3/hour has been reported as low as 0.43 US$/1000 gal corresponding to
0.11US$/m3. [255] The choice of 2.3 US$/m3 reflects the price of the water itself, the
demineralization and perhaps a further purification in order to avoid problems with
impurities polluting the SOEC.
4.5.3.1 Calculated H2 production cost
This section presents H2 cost calculations based on the assumptions in Table 4-3.
Whenever the assumptions differ from the ones in the table it is specified below the
figure.
88
Figure 4-24. H2 production cost vs. electricity price at various investment costs.
Details on the assumptions for the calculation are specified in Table 4-3. The pie
diagram show the parts of the production price.
Figure 4-24 shows the hydrogen production price as function of electricity price. The
estimation is based on the assumptions in Table 4-3 and the cell performance
measured in the i-U curve shown in Figure 4-23. Details on the calculation methods
are given in Appendix. The point marks the H2 cost given the assumptions in Table
4-3. In this case the calculated current density was -1.5 A/cm2 and the H2 cost was 71
US¢/kg equivalent to 30$/barrel crude oil using the HHV. Note that electricity
accounts for the main part of the production cost and that the second major
contribution to the production cost is to pay off the loans.
Figure 4-25. H2 cost vs. lifetime and investment. Details on the calculation
assumptions are specified in Table 4-3. The pie diagram show the parts of the
production price.
Figure 4-25 shows the H2 cost vs. the life time at different investment costs. Details
on the economy assumptions are given in Table 4-3 and the calculation method is
discussed in Appendix 2 in this chapter. Note that the lifetime can be reduced to about
5 years without dramatic increases in production cost.
89
Figure 4-26. H2 cost vs. cell voltage calculated various investment costs and
lifetimes. Details on the assumptions for the calculation are specified in Table 4-3.
Figure 4-26 shows the H2 cost vs. cell voltage at various investment costs and
lifetimes. It is interesting to note how the economic balance between high production
rates (high cell voltage) and high efficiency (low cell voltage) changes with the
lifetime and investment cost: At 5 years lifetime and >4000 $/m2 it is most favourable
with a high production rate and a low efficiency. At 10 years lifetime and <4000$m2 it
is most favourable with a high efficiency and a lower production rate.
Figure 4-27 shows H2 cost vs. cell voltage at various operation temperatures and
investment costs. Details on assumptions for the calculation are specified in Table
4-3.. Again it is interesting to see how the economic balance between high production
rates and high efficiencies changes with cell temperature (i.e. internal resistance and
in turn current density) and investment costs. In general a high cell temperature and a
low investment cost shift the economic balance towards higher production
efficiencies.
90
Figure 4-27. H2 cost vs. cell voltage at various operation temperatures and investment
costs. Details on assumptions for the calculation are specified in Table 4-3.
Figure 4-28 shows the H2 cost vs. pressure. The reason why the production costs
increases with pressure is because the current density at a constant cell voltage
decreases with increasing pressure due to the increase in the reversible voltage (i.e.
cell voltage at 0 A/cm2). It should be noted that the volumetric energy density is
higher in pressurized H2 than in un-pressurized H2 which makes it more valuable due
to a more efficient storage and handling.
91
Figure 4-28. H2 cost vs. pressure and investment cost. The cell voltage was 1.31 V
and the heat reservoir was 10 °C warmer than the H2O evaporation temperature.
Further details on the assumptions for the calculation are specified in Table 4-3.
4.5.3.2 Calculated CO production costs
This section presents CO cost calculations based on the assumptions in Table 4-3.
Whenever the assumptions differ from the ones in the table it is specified below the
figure.
It is difficult to get a good estimate of what CO2 would cost if it is based on a
non- fossil fuel technology. The Gibbs free energy required to capture CO2 from air is
140 kWh/ton. If electricity at a cost of 1.3 US¢ is used, this corresponds to 1.8
US$/ton. Since technology for such a "wind scrubber" technology is very immature
no good price estimation could be found.
The historic stock market price for CO2 emission allowances is presented in
Figure 4-29. Note that the final low price was due to an over-allocation of allowances
to the market. The stock market price for EUA 2008 (CO2 allowances in 2008) are
currently about 15 €/ton CO2.
Figure 4-29. The graph show the development of Point Carbon's bid-offer closing
price for EU allowances (EUA 2005, EUA 2006 and EUA 2007).
Since the CO is produced by means of electrolysis of CO2 there is no overall CO2
emission when burning off the CO. Hence the CO should be economically
competitive with fossil CO (i.e. CO made from coal) if the production cost equals that
of fossil CO plus the cost of CO2 emission allowance.
Figure 4-30 shows the CO production cost as function of electricity price. The
estimation is based on the assumptions in Table 4-3 and the cell performance
measured in the i-U curve shown in Figure 4-23. Details on the calculation methods
are given in Appendix. The point marks the CO cost given the assumptions in Table
4-3. In this case the calculated current density was -1.6 A/cm2 and the CO cost was
5.6 US¢/kg equivalent to 34$/barrel crude oil using the HHV. Note how the electricity
accounts for the main part of the production cost and that the second major
contribution to the production cost is to pay of the loans.
Carbon formation could probably cause the Ni/YSZ electrode to degrade and
this issue has to be investigated. However, the electrode seemed stable towards carbon
formation at the maximum cell voltage in the CO2 i-U curve in Figure 4-23. Here the
maximum current density was -1.5 A/cm2 at 950 °C and 1.29 V which is not far from
the -1.6A/cm2 that is used in the calculation.
92
Figure 4-30. CO cost vs. electricity price at various investment costs. The cell voltage
was 1.47 V. Details on calculation assumptions are specified in Table 4-3.
Figure 4-31 shows the CO cost vs. lifetime. At 6 years the electricity still constitutes
the major part of the CO cost.
Figure 4-31. CO cost vs. lifetime. The cell voltage was 1.47 V. Details on calculation
assumptions are specified in Table 4-3.
Figure 4-32 shows the CO cost vs. cell voltage. At low investment cost it is preferable
to operate the cell at a low cell voltage, and Figure 4-33 gives the dependence on
pressure.
93
Figure 4-32. CO cost vs. cell voltage at various investment costs. Further details on
the assumptions for the calculation are specified in Table 4-3.
Figure 4-33. CO cost vs. pressure. The cell voltage was 1.47 V. Further details on the
assumptions for the calculation are specified in Table 4-3.
4.5.4 Discussion of the results of the economic calculations
It is seen from the pie diagrams in Figure 4-24, Figure 4-25, Figure 4-30 and Figure
4-31 that the electricity costs dominates the investment of the instalment for both the
H2 production costs. This incident combined with the fact the H2O and CO2
electrolysis reaction is highly endothermic provides a unique opportunity to make an
energy conversion that is both economically viable and highly efficient. From Figure
4-26 and Figure 4-27 it is seen that this situation is “realized” when the investment
costs are low, the lifetime is long and the operating temperature is high (i.e. the
internal resistance is low).
The major obstacle for this technology is the lifetime which has to be
improved. During the last three years it has been shown on Risø that the
stoichiometric albeit glass sealings that usually are used to seal the gasses from the
surroundings seem to make the Ni/YSZ electrode degrade with unacceptably short
lifetimes as result. A seal-less test of a Risø cell at Karlsruhe/EIfER strengthen this
hypothesis. The test was performed at 800 °C, -0.5A/cm2, and 50% H2O in the inlet
gas. The cell voltage increase was 15 mV/1000h. This is still too high to achieve the
desired lifetime of 5 to 10 years.
It is suggested that further R&D should focus on lowering the investment costs
to app. 2000$/m2 and to assure lifetimes of more than 5 years while keeping the
realized performance, corresponding to current densities above 1 A/cm2 at thermoneutral potential. This will enable the technology to produce H2 and CO having the
lowest production costs around the thermo-neutral potential. At these cell voltages the
“electricity to fuel” efficiency is more than 90% for both H2 and CO production.
It should be noted that a more detailed calculation should incorporate the loss of heat
to the surroundings which will decrease the overall efficiency. However these losses
94
can be minimized using inexpensive materials such as mineral wool and by using
large scale electrolyser units.
A 10 l pressure bottle, capable of 230 bar cost app. 270 $. [256,257] The
SOEC stack density is app. 0.25 m2 cell area per litre. [258] Hence, the expense to
pressurization can be approximated to be
270 kr
l
108$
⋅
= 2
2
10 l
0.25 m
m
(29)
Compared with the 2000-6000 $/m2 installation expense used in the economic
assumption, the expense to pressurization facilities is believed to be small.
4.5.4.1 Synthetic fuel production
Methanol and DME
A Cu/ZnO-Al2O3 catalyst is normally used for methanol synthesis with operation
conditions around 200 °C - 300 °C at 4.5-6 MPa. [259,260] At 260 °C the steam
pressure is 4.5 MPa and at 275 °C it is 6 MPa. At such operating conditions the C to
CH3OH conversion ratio is limited (< 20%). [259] Also dimethylether (DME) can be
produced in one step using similar catalysts at somewhat modified conditions.
In order to achieve acceptable yields recycling could be used as well has a
higher pressure at the catalyst. This will, however, reduce the current density in the
SOEC stack which will result in a higher H2 and CO cost as seen in Figure 4-28 and
Figure 4-33. It may also be possible to pressurize only the exhaust gasses from the
SOEC stack in order to keep the high SOEC performance.
Experiments with pressurised SOEC operation are limited in the literature. A
project in pressurised SOEC operation will be initiated at Risø National Laboratory
during 2007.
Methane
Pipe lines and storage facilities for natural gas are widely established in Denmark. In
this perspective, it is interesting to produce methane from renewable energy sources
to feed in to the natural gas pipe lines. Ni -based catalysts are typically used for
methanation. Typical operating conditions are 190 °C – 450 °C, typically at lower
pressures than methanol catalyst operating pressures. Hence it should be possible to
reach an acceptable yield in one cycle - even without pressure differentiation in the
heat exchanger.
Hydrogen
Heat producing power plants such as nuclear and geothermal can be used for steam
generation and pressurizing (super heated steam) which avoid the need for pumps.
[209,227,235,261,262] The Danish natural gas pipe lines are capable of handling 5%
- 10% hydrogen without significant changes. [263]Large storage facilities for natural
gas, such as the ones in Ll. Thorup and Stenlille, are already established in Denmark.
95
4.5.4.2
Carbon Dioxide Neutral Synthetic Fuels by Capture of
CO2 from the Atmosphere
Within recent years a huge rise in the number of abnormal weather events has
occurred. Meteorologists agree that these exceptional conditions are signs of a Global
Climate Change. Scientists agree that the most likely cause of the changes is manmade emissions of the so-called greenhouse gases that trap heat in the earth's
atmosphere. Although there are six major groups of gases that contribute to the global
climate change, the most common is carbon dioxide (CO2). For this reason there are
much research in sequestration of CO2 from power plants and other point sources for
storage and removal of CO2. As mentioned in the perspectives for SOEC (this report),
synthesis gas and thereby synthetic fuel can be produced by electrolysis of CO2 in fuel
cells. CO2 capture from air and recycling or reuse of CO2 from energy systems would
therefore be an attractive alternative to storage of CO2 in the underground and would
provide CO2 neutral synthetic hydrocarbon fuels.
Capture of CO2 for recycling can be achieved by absorption processes
employing amines or carbonates as absorbents. The regeneration includes heating of
the absorbent; therefore reduction of the energy requirement becomes a determining
factor for realizing CO2 recycling. From the viewpoint of energy saving in
regeneration of the absorbent, carbonates are preferable to amine solutions, since the
energy requirement for CO2 removal in the carbonate process is about half of that of
the amine process [264]. However the rate for CO2 absorption and desorption with
carbonates is slow, but for CO2 capture/recycling from air, the absorption and
desorption rate may not be a determining factor. A carbon neutral energy cycle
utilizing CO2 capture from air with calcium carbonate in combination with a fuel cell
is sketched in Figure 4-34.
CO2
into the
atmosphere
Power
H2O into the
atmosphere
Consumption: Fuel cell
or Otto engine
Ca(OH)2 + CO2 (from air) Æ
CaCO3 + H2O
Transport
CaO + H2O Æ
Production of fuel:
Ca(OH)2 + heat
Catalysis:
CH4 or CH3OH
Electrolysis: CO + H2
CaCO3 + heat Æ
CaO + CO2
CO2
H2O
Electricity
from wind or
water
Figure 4-34. Carbon dioxide neutral energy cycle utilizing CO2 capture from air with
calcium carbonate in combination with a Solid Oxide Electrolysis Cells (SOEC).
Mineral carbonation has been recognized as a potentially promising route for
permanent and safe storage of carbon dioxide, and thereby also a promising route for
recycling of CO2. Both the potentially large CO2 sequestration capacity and the
96
exothermic nature of the carbonation reactions involved have contributed to an
increasing amount of research on mineral carbonation in recent years [265,266]. A
number of different carbonation process has been reported, of which aqueous mineral
carbonation route was selected as the most promising in a recent review [265].
Calcium carbonate is a well known CO2 absorbent [265]. Also the less known
magnesium carbonate can be employed. As previous mentioned the required energy
for regeneration is a determining factor, therefore equilibrium and reaction enthalpies
of CO2 recycling utilising either calcium carbonate (CaCO3, limestone), magnesium
carbonate (MgCO3, magnesite) and a mixture of calcium carbonate and magnesium
carbonate (CaMgCO3, dolomite), of which calcium carbonate and calcium
magnesium carbonate are abundant present in nature, are calculated in order to
evaluate the feasibility to recycle carbon dioxide. All equilibrium and reaction
enthalpies were calculated using Factsage 5.5 Software [267]and are all calculated at
1 atmosphere pressure.
Calcium Carbonate Cycle
Calcium carbonate can be formed by passing carbon dioxide into a solution of
calcium hydroxide (Ksp for Ca ( OH)2 ∼ 4.68·10−6 at 298K):
Ca(OH)2 + CO2 → CaCO3 + H2O
(30)
Reaction (30) is exothermic at all temperatures as can be seen from Figure 4-35.
Calcium carbonate is poorly soluble in water. Consequently calcium carbonate
precipitates out of the solution, driving the reaction towards the formation of CaCO3.
50
Energy (KJ)
0
T ΔS
-50
ΔG
-100
ΔH
-150
0
100
200
300
400
500
600
700
800
900
1000
Temperature (ºC)
Figure 4-35. Thermodynamics of calcium hydroxide and carbon dioxide mixing.
Solid calcium carbonate releases carbon dioxide on heating, to form calcium oxide
according to equation (31). Calcium carbonate exists in equilibrium with calcium
oxide and carbon dioxide at any temperature. At room temperature the equilibrium
favours calcium carbonate, because the equilibrium CO2 pressure is only a small
fraction of the partial CO2 pressure in air. At temperatures above 550 °C the CO2
pressure begins to exceed the partial pressure of CO2 in air as shown in Figure 4-36A.
Thermodynamically the CO2 release (negative Gibbs free energy for reaction (31)
occurs above 880 ºC as can be seen from Figure 4-36B.
CaCO3 → CaO + CO2
(31)
97
A
B
9
80
8
60
7
40
Energy (KJ)
6
5
4
20
-40
-60
1
-80
0
-100
600
700
800
CO2 release
-20
2
500
CO2 uptake
0
3
0
900
100
200
300
400
500
600
700
800
900
Temperature (ºC)
Temperature (ºC)
Figure 4-36. Thermodynamics of carbon dioxide release/uptake from calcium
carbonate.
The formed calcium oxide reacts vigorously with water to form calcium hydroxide, as
in the following equation:
CaO + H2O → Ca(OH)2
(32)
Gibbs free energy
100
80
60
40
Energy (KJ)
Vapour pressure CO2
Gibbs free energy
100
20
H2O release
0
H2O uptake
-20
-40
-60
-80
-100
0
100
200
300
400
500
600
700
800
900
1000
Temperature (ºC)
Figure 4-37. Thermodynamics of water release/uptake from calcium oxide.
The combination of reaction (30) to (32) closes the CO2/carbonate cycle for CO2
capture/recycling as shown in Figure 4-34 and Figure 4-38. From Figure 4-38 it can
be seen that a temperature interval between 500 and 880 ºC is needed to perform CO2
capture/recycling with calcium carbonate.
98
1000
Gibbs free energy
100
CaO
1
CaCO3 → CaO + CO2
80
Composition (mole)
60
Ca(OH)2
Energy (KJ)
40
CO2
CaCO3
0
-20
Recycling temperature
CaO + H 2O → Ca(OH)2
-40
CaCO3 →
Ca(OH)2 → CaO + H 2O
20
-60
CaO + CO2
-80
CaCO3
0
0
100
200
300
400
500
600
700
800
900
-100
0
1000
100
200
300
400
500
600
700
800
900
Temperature (ºC)
Temperature (ºC)
Figure 4-38. Composition upon heating after mixing Ca(OH)2 + 0.50 CO2 + H2O
(initially all CO2 is consumed to form CaCO3).
As mentioned above also magnesium carbonate can be used for CO2 capture/recycling
as will be described below.
Magnesium Carbonate
As for calcium carbonate, magnesium carbonate can be formed by passing carbon
dioxide into a solution of its hydroxide. Because magnesium carbonate is only slightly
soluble in water, magnesium carbonate precipitates out of the solution:
Mg(OH)2 + CO2 → MgCO3 + H2O
(33)
Magnesium carbonate releases carbon dioxide on heating, to form magnesium oxide
according to equation (34). Compared to calcium hydroxide, magnesium hydroxide
releases CO2 at a much lover temperature. The Gibbs free energy for the release of
CO2 becomes negative at 400 ºC as can be seen from Figure 4-39. Further it can be
seen from Figure 4-41 that starting at around 275 ºC, MgCO3 releases CO2 because of
the equilibrium with CO2 in air.
MgCO3 → MgO + CO2
(34)
Gibbs free energy
100
80
60
Energy (KJ)
40
20
CO2 uptake
0
CO2 release
-20
-40
-60
-80
-100
0
100
200
300
400
500
600
700
800
900
1000
Temperature (ºC)
Figure 4-39. Carbon dioxide release from magnesium carbonate as a function of
temperature.
99
1000
At temperatures below 270 ºC, magnesium oxide reacts with water to form
magnesium hydroxide. Above 270 ºC water is released forming magnesium oxide
from magnesium hydroxide:
MgO + H2O → Mg(OH)2
(35)
Gibbs free energy
100
80
60
Energy (KJ)
40
20
H2O release
0
H2O uptake
-20
-40
-60
-80
-100
0
100
200
300
400
500
600
700
800
900
1000
Temperature (ºC)
Figure 4-40. Formation of magnesium hydroxide from magnesium oxide and water.
As a consequence of the lower reaction temperature for both reaction (34) and (35), a
carbonate cycle operating with magnesium carbonate can be operated at temperature
cycles between 250 ºC and 400 ºC as shown in Figure 4-41.
Gibbs free energy
100
1
MgO + H2O → Mg(OH)2
80
MgO
Composition (mole)
60
MgCO3
CO2
Mg(OH)2
Recycling temperature
MgCO3 → MgO + CO2
-80
Mg(OH)2
0
100
0
-20
-60
MgO + H2O
0
20
-40
Mg(CO3)2 → MgO + CO2
Mg(OH)2 →
Energy (KJ)
40
200
300
400
500
600
700
800
900
Temperature (ºC)
1000
-100
0
100
200
300
400
500
600
700
800
900
Temperature (ºC)
Figure 4-41. Composition upon heating after mixing Mg(OH)2 + 0.25 CO2 +
H2O(initially all CO2 is consumed to form MgCO3).
Calcium Magnesium Carbonate Cycle
In nature magnesium carbonate is found as calcium magnesium carbonate (Dolomite,
CaMg(CO3)2. The CO2 release upon heating of solid calcium magnesium carbonate
occurs in two steps corresponding to the decomposition of MgCO3 or CaCO3 as
shown in Figure 4-42:
100
1000
Composition (mole)
2
CO2
CaMg(CO3)2
1
MgO
CaO
CaCO3
CO2, MgO
CaMg(CO3)2 →
CaCO3 →
MgO + CO2 + CaCO3
CaO + CO2
CaCO3
MgCO3, CaO
0
0
100
200
300
400
500
600
700
800
900
1000
Temperature (ºC)
Figure 4-42. Composition upon heating after mixing Mg(OH)2 + Ca(OH)2 and 2 CO2
(initially all CO2 is consumed to form CaMg(CO3)2)
Also the reaction between the oxide and water also occurs in two steps corresponding
to the reaction with either MgO or CaO:
Mg(OH)2,
Composition (mole)
1
Ca(OH)2
Mg(OH)2 →
Ca(OH)2 →
MgO + H2O
CaO + H2O
MgO, CaO
0
0
MgO, CaO
MgO
100
200
Ca(OH)2
CaO, Mg(OH2)
300
400
500
600
700
800
900
1000
Temperature (ºC)
Figure 4-43. Formation of calcium magnesium hydroxide from calcium oxide,
magnesium oxide and water.
As a consequence of the lower reaction temperature for CO2 release from magnesium
carbonate and the lower reaction temperature for the decomposition of magnesium
hydroxide to magnesium oxide, a carbonate cycle with calcium magnesium carbonate
can be operated between 250 ºC and 400 ºC utilizing magnesium carbonate only.
Since only half of the mineral is active, this would result in a higher amount of
minerals needed to be transported, and more rock would have to be mined. More work
on the kinetic of the adsorption and desorption processes as well as the practical
process implementation is necessary to estimate the economic aspects regarding CO2
recycling for the production of CO2 neutral synthetic fuels.
Energy consumption by capture of carbon dioxide
The energy consumption by capture of carbon dioxide from the atmosphere where the
CO2 concentration currently is 360 ppm is only 5.6% compared to the combustion of
carbon to give one mol CO2. Having in mind that the free energy of the capture (in the
case of full reversibility) is the same as ΔGMix of CO2 of 1 atm with atmospheric air,
the calculation was done as follows:
101
1 mol CO 2 = 3.6 × 10-4 atm (PCO2 in atmosphere)
x mol Air = 1 - 3.6 × 10-4 atm
mol CO 2 =
V=
V=
PAir ⋅ V
R ⋅T
n ⋅R ⋅T
V = Air
PAir
PCO2 ⋅ V
mol Air =
R ⋅T
n CO2 ⋅ R ⋅ T
PCO2
n CO2 ⋅ R ⋅ T
PCO2
=
n Air ⋅ R ⋅ T
PAir
n Air
1
=
-4
3.6 × 10
1 - 3.6 × 10-4
1 - 3.6 × 10-4
n Air =
3.6 × 10-4
1 - 3.6 × 10-4
n Total = n CO2 + n Air = 1 +
3.6 × 10-4
ΔG Mix = n Total ⋅ R ⋅ T ⋅ x CO2 ⋅ ln CO2 + x Air ⋅ ln Air
V=
(
)
⎛
1 - 3.6 × 10-4 ⎞
-4
-4
-4
-4
ΔG Mix = ⎜ 1 +
⎟ ⋅ 8.3145 ⋅ 298.15 ⋅ 3.6 × 10 ⋅ ln ( 3.6 × 10 ) + (1 − 3.6 × 10 ) ⋅ ln (1 − 3.6 × 10 )
-4
3.6 × 10 ⎠
⎝
ΔG Mix = - 22.1353 KJ
molCO2
(
The energy by combustion of carbon to give one mol CO2:
C + O 2 → CO 2
ΔH f CO2 = -393.51 KJ
molCO2
ΔG Mix for release of CO2 is 5.6% of ΔH for carbon combustion.
4.5.5 Conclusions on SOEC
It is found that, given the assumptions in Table 2, H2 and CO can be produced at
attractive production costs, using SOECs. The H2 production cost was found to be 71
US¢/kg equivalent to 30 $/barrel crude oil using the HHV. The CO production cost
was found to be 5.6 US¢/kg equivalent to 34 $/barrel crude oil using the HHV.
If heat for steam generation can be provided from a waste heat source, the
production price can be lowered even further. Electrolysis on a H2O/CO2 mixture will
produce synthesis gas which can be catalyzed into various types of synthetic fuels. In
such a synthetic fuel production, some reduction in the production price may be
achieved by utilizing the heat from the catalysis reaction for steam generation. The
main part of the production cost for both H2 and CO is the electricity cost.
For lifetimes above 3-4 years the H2 production price starts to become
insensitive to the life time. For the CO production price this is about 6 years. The
production cost was found to be lowest at ETn. Here the efficiency from electricity to
fuel was found to be 93% for CO production and 96% for H2 production. These
figures do not include heat loss to the surroundings.
102
)
In a combined CO2/H2O electrolysis operation, an additive of Cu to the
Ni/YSZ support layer may enhance the CO production due to a WGSR in the support
layer combined with electrolysis in the electrode.
Recycling of CO2 for carbon dioxide neutral synthetic fuels can be performed
with calcium carbonate, which is a well-known CO2 absorbent and has been
suggested for CO2 sequestration. Thermodynamic equilibrium calculations show on
the other hand that CO2 capture/recycling using magnesium carbonate can be operated
at approximately 400 ºC lower than the 800 ºC for calcium carbonate.
Magnesium carbonate is abundant in nature as calcium magnesium carbonate.
A carbonate cycle for CO2 capture with calcium magnesium carbonate can be
operated between 250 ºC and 400 ºC utilising magnesium carbonate only. Using only
magnesium carbonate from calcium magnesium carbonate, higher amount of minerals
would have to be mined and transported.
A carbonate cycle for CO2 capture/recycling is definitively technically
feasible, but the practical and economic aspects regarding calcium carbonate,
magnesium carbonate or calcium magnesium carbonate have to be assessed to
determine the most suitable absorbent for CO2 capture/recycling.
In total, these findings seem quite interesting for a further investigation in hydrogen
and synthesis gas production by use SOEC technology.
Some main subjects to be addressed in the future R&D are: 1) precise
identification of the mechanism of the cell degradation, 2) developments of highly
durable cells, 3) further feasibility studies through cell and stack testing, 4)
construction of pressurized cell and stack test facilities, 5) construction of prototype
electrolyzer systems, 6) more detailed technical and economical modelling should be
done parallel to the experimental work.
4.5.6 Appendix 1. CCASR calculation
The following input parameters are used to calculate CCASR: The cell voltage U, the
current through the cell I, the cell area A of one cell in the stack and the H2 and H2O
gas flow to the Ni/YSZ electrode of the cell and the O2 and/or air flow to the
LSM/YSZ electrode.
Since the stoichiometry and number of electrons involved in the H2O and CO2
electrolysis reaction is identical, the calculation presented below is also valid for CO2
electrolysis except that the thermo-neutral voltage ETn is different for CO2
electrolysis.
The cell is assumed to be divided into 10 slaps. The resistance of each slap is
taken to be 10 times the resistance of the whole cell. The gas flow is assumed to be a
co-flow. 1
The CCASR is found by iteration. First the Nernst voltage of the H2/H2O inlet gas vs.
O2 is calculated. This is labelled ENernst-1. An initial guess on CCASR is labelled
CCASR(itt.1). The current density through the first slap of the cell is now estimated
as
ENernst-1 − U
= i1
CCASR(itt.1)
(36)
1
In the cell test setup the gas flow is a cross flow. In these experiments the LSM/YSZ gas flow is pure
O2. This means that the oxygen partial pressure remain constant throughout the whole electrode. Hence
it does not affect the calculations whether a cross flow or a co-flow calculation is used.
103
A i1
⋅
from the H2 flow
10 2 F
rate at the first slap. The H2O flow rate at the second slap is found by adding
A i1
⋅
to the flow rate at the first slap. ENernst-2 is found for the second slap using the
10 2 F
flow rates at the second slap as partial pressures in the Nernst equation and i2 is found
by substituting ENernst-2 into (36). The current through the cell is found as
The H2 flow rate to the second slap is found by subtracting
(i1 + i2 + .. + i10 )A
= I (itt.1)
10
(37)
CCASR(itt.2) is found as
CCASR(itt.2) = CCASR(itt.1) −
I − I (itt.1)
N
(38)
N is a large positive constant used to assure that the iteration converge towards a
stable value. CCASR(itt.2) is inserted into (36) and step (36) to (38) is repeated. The
iteration stops when I(itt.n)=I. CCASR is then given as CCASR(itt.n).
4.5.7 Appendix 2. Economy calculation
The calculation method is based on the data in Table 4-3. The calculation describes
steam electrolysis but since the stoichiometry and number of electrons involved in the
H2O and the CO2 electrolysis reaction are identical, the calculation is also valid for
CO2 electrolysis with only minor changes. The differences are specified whenever
they occur.
Using the cell temperature given in Table 4-3, the average CCASR is found
from the third order polynomial shown in Figure 4-23. This value is used as
CCASR(itt.1) in equation (36) and given the cell voltage, inlet gas concentrations,
operation pressure and temperature from Table 4-3 we can find i1. An initial guess on
the H2O flow rate to one cell is applied to find i2…i10 and the total current I as in
equation (37). Subsequently the H2O flow rate is found by iteration using equation
(39).
X& H2O ( itt.2 )
⎛
⎞
I ( itt.1)
⎜⎜ 0.95 −
⎟
2 F ⋅ X& H2O ( itt.1) ⎟⎠
⎝
= X& H2O ( itt.1) +
N
(39)
where X& H2O ( itt.1) is the initial guess on the H2O flow rate to one cell given in mole
per second. N is a large constant to assure the iteration converge to a stable value. The
104
iteration stops when X& H2O ( itt.( n + 1) ) = X& H2O ( itt. n ) and the H2O utilization is 95%.
The current through the cell now given as I ( itt.n ) , also found from equation (39).
The electric power consumption P for the SOEC stack is given as
P = − I ⋅U ⋅ y
(40)
where y is the number of cells in the stack. The regular outgoing for electricity is
found as
YElectricity = P ⋅ Electricity price ⋅ Operating activity
(41)
The regular outgoing for heat from the heat reservoir is found as
YHeat reservoir =
( ΔH (T ) − ΔH (T ) ) ⋅ X& ( itt. n ) ⋅ y ⋅ Heat price ⋅ Operating activity
o
Gas
H.r.
o
Gas
r
(42)
H2O
o
o
where ( ΔH Gas
(TH.r. ) − ΔH Gas
(Tr ) ) is the enthalpy change, i.e. the energy, required to
heat the H2O from room temperature to the temperature of the heat reservoir. Note
that since the heat reservoir temperature is 110 °C and the operating pressure is 0.1
MPa, the H2O evaporates at the heat reservoir and the heat required for the
evaporation is provided by the heat reservoir.
The energy loss rate in the heat exchanger calculated as
o
o
Inlet
Outlet
H.E. loss rate = 5% ⋅ ∑ ( ΔH Gas
(T ) − ΔH Gas
(TH.r. ) ) ( X& Gas
+ X& Gas
)
Gas
(43)
o
o
where ΔH Gas
(T ) − ΔH Gas
(TH.r. ) is the energy required to heat the gas specie from the
Inlet
Outlet
and X& Gas
is the
heat reservoir temperature to the stack operation temperature. X& Gas
flow rates of the gas specie in question in the inlet and outlet part of the heat
exchanger respectively. In order to avoid the cell from cooling down, the H.E. loss
rate has to be compensated by heat at the stack operating temperature. This can be
done by an electric heater or (preferably) by keeping the cell voltage slightly higher
than the thermo-neutral voltage.
The regular outgoing to the loss in the heat exchanger is calculated as
YHeat exhanger = H.E. loss rate ⋅ Electricity price ⋅ Operating activity
(44)
The regular outgoing to demineralised water is given as
YDemineralized water = X& HInlet
⋅ Demineralised water cost ⋅ Operating activity
2O
(45)
105
For CO2 electrolysis the expenses for water is substituted with the CO2 cost specified
in Table 4-3.
Finally the regular outgoing to pay off the loan is calculated as an annuity loan
based on the investment, the interest rate and life time.
.
The total regular outgoing is given as
YTotal = YElectricity + YHeat reservoir + YHeat exchanger + YDemineralized water + YLoan
(46)
The H2 production cost finally is given as
H 2 production cost =
YTotal
2F ⋅ YTotal
=
H 2 production rate
-I
(47)
106
5 Commercially available electrolyzers
The technology and sizes of commercially available electrolysers vary greatly. In this
chapter we will only focus on technology that can be useful to the electrical power
grid.
5.1 Performance review
An electrolyzer is an on-site hydrogen generating plant based on water electrolysis.
Electrolysis takes place when an electric current flows through an electrolyte (in this
case, water) from anode to cathode. Water molecules are spontaneously split into
hydrogen and oxygen gases of high purity, and the resulting gases can then be
purified, compressed, stored or distributed according to the requirements.
Three types of electrolyzers can be considered for the production of hydrogen
by use of wind power for grid balancing and production of fuels for transport. Solid
Oxide Electrolyzers (SOE), Proton Exchange Membrane Electrolyzer (PEME) and
Alkaline Electrolyzers (AE). The figure below shows where the three technologies are
located on the road from research through development to commercial products.
Research - Product development - Commercial
products
Solid Oxide Electrolyses
Proton Exchange electrolyses
Alkaline Electrolyses
Figure 5-1. State of development of the different electrolyzer types.
107
Solid Oxide Electrolysers (SOE)
Figure 5-2. Solid oxide stack.
SOE is based on a high temperature technology used in Solid Oxide Fuel Cells
developed by RISØ. The technology offers the possibility of very high efficiencies of
more than 90 %.
SOE is in the very beginning of the research phase. Extremely high current
densities have been showed in the laboratory, but there is still much work to be done
in order to develop a prototype.
Because SOE is not expected to be commercial viable within the next 3 year
this technology will not be included in the market survey.
Proton Exchange Membrane Electrolyser (PEME)
Figure 5-3. Two PEM electrolyzer units.
108
PEME is commercial in sizes op to 44 kW. They deliver hydrogen of high purity and
are suitable for pressurizing. Their power density is quite high and therefore useful
where space is limited and expensive. As for PEM fuel cells the lifetime is still not
sufficient for all applications. More than 40,000 hours is not to be expected.
The most common applications are in labs, submarines and spacecrafts. The
efficiency is in the lower end and the price in the high end.
Because of the relative small maximum capacity at present, the limited lifetime and
the high price, the PEME will not be included in this market survey.
Alkaline Electrolysers (AE)
Figure 5-4. Large scale alkaline electrolyzers
Production of hydrogen by alkaline electrolysers is an about 100 year old technology
used in the chemical and metallurgic industry and for production of fertilizer in the
form of ammonia NH3. The energetic efficiency on converting electricity to hydrogen
is reasonably high on modern plants: between 80 and 90%. The lifetime is as high as
20 years, with a major service check every 6 years.
Electrolyser marked, Western and Eastern
The relative high prices of electrolysers are a barrier for the use of low price wind
power for hydrogen production. It is also a significant barrier for the application of
hydrogen as an energy carrier in the transport sector and as fuel for micro CHP
systems, applications there are a growing interest for in EU, Japan and USA.
Therefore there is no doubt there will be an incising market for electrolysers if cheap
electrolysers can be brought to the market.
109
In Eastern Europe, Poland, Ukraine and especially Russia there are through
out the years developed knowledge, experience and expertise concerning electrolyser
plants among other to be used in furnaces and ovens in the metal industry and for
cooling of power plant generators. However, these companies are not at the moment
able to deliver plants to the western market because of the requirements to guaranties,
CE-marking, service, etc.
5.1.1 Western marked
Five major western suppliers of alkaline electrolysers have been identified as:
•
•
•
•
•
Hydro in Norway
Hydrogenics in Belgium
Iht in Switzerland
AccaGen in Switzerland
Erre Due in Italy represented by H2Indistrial in Denmark
The plants can be divided in two groups. Atmospheric and pressurized plants. The
atmospheric plant operates at atmospheric pressure of one bar and the pressurized
plants operate at pressures from 4 to 30 bar depending of the make.
Table 5-1. Manufacturers of AEC stacks and their power ratings.
Western
Supplier
Hydro
Hydrogenics
Iht
AccaGen
Erre Due
Atmospheric
plants
200 to 2000 kW
14 to 1500 kW
Pressure
plants
50 to 300 kW
60 to 240 kW
500 to 3400 kW
7 to 500 kW
100 to 200 kW
Below is showed the filled in questionnaire from the five manufactures. It has not
been possible to get all information from all manufactures.
110
5.1.1.1 Norsk Hydro Electrolysers AS
Heddalsveien 11
P.O. Box 44
N-3671 Notodden
Norway
Phone+47 35 09 39 99
Fax+47 35 01 44 04
E-mail [email protected]
Personal contacts:
Marketing and Sales Director
Mr. Roy Grelland
[email protected]
R&D Manager, Hydro Technology Ventures
Dag Øvrebo
Managing Director, Hydro Electrolysers
Knut Harg
Figure 5-5. Alkaline Electrolyzers from Norsk Hydro.
111
Table 5-2. Norsk Hydro electrolyzers (ambient pressure)
Specifications of Electrolysers
(no external equipment, gas
separator, cooling
pumps, power supply, etc.)
Max capacity
Max power
Efficiency
Number of cells
DC Voltage
DC Current
Current density
Outlet Pressure
Operation range
Operation speed
Purity
Weight
Diameter
Circulation of electrolyte
KOH
Lifetime
Operation temperature
Maintenance costs
Service time
Anode material
Cathode material
Diaphragm material
Price
Price
Price
Hydro
1 Bar
name
/model
name
/model
name
/model
name
/model
Nm3/h
kW
Kwh / Nm3 H2
5010
50
205
4.1
31
5020
150
615
4.1
92
5030
300
1230
4.1
183
5040
377
1546
4.1
230
4000
4000
4000
4000
1
20 - 100
12
minutes
99.9
1
20 - 100
1
20 - 100
12
12 minutes minutes
99,9
99,9
1
20 - 100
12
minutes
99,9
80
80
80
V
Amp
A /cm2
Bar
from x % to y%
dA / dt
% hydrogen
Kg
m
l/min
%
Years
Celsius
USD/ year
Years
80
Euro
USD
DKK
112
Table 5-3. Norsk Hydro electrolyzers (elevated pressure)
na
me
mo
del
na
me
mo
del
na
me
mo
del
na
me
mo
del
na
me
mo
del
na
Hydr me
mo
Specifications
of o
Electrolysers
12 Bar del
(no
external
equipment,
gas
separator, cooling
pumps, power supply,
etc.)
Max capacity
Nm3/h 10
na
me
mo
del
na
me
mo
del
na
me
mo
del
na
me
mo
del
12
16
77
20
96
24 30 40 50 60 65
115 144 192 240 288 312
Max power
kW
48
Kwh /
Nm3
H2
4.8
58
4.8
4.8
4.8
DC Voltage
V
60
72
DC Current
Amp
A
/cm2
Bar
from x
% to
y%
dA / dt
%
hydrog
en
Kg
m
820 820
Efficiency
Number of cells
4.8
4.8
4.8
4.8
4.8
4.8
104 128 160 192 136 168 200 216
Current density
Outlet Pressure
Operation range
Operation speed
Purity
Weight
Diameter
Circulation
of
electrolyte
KOH
Lifetime
Operation temperature
Maintenance costs
Service time
Anode material
Cathode material
Diaphragm material
Price
Price
Price
760 760 760 760 150 150 150 150
0
0
0
0
12 12 12 12 12 12 12 12 12 12
50 50 50 - 50 - 50 - 50 - 50 - 50 - 50 - 50
100 100 100 100 100 100 100 100
100 100
99.
99. 99. 9
9
9
99.
9
99.
9
99.
9
99.
9
99.
9
99.
9
l/min
%
Years
Celsius
USD/
year
Years
Euro
USD
DKK
113
99.
9
5.1.1.2 Hydrogenics Europe N.V
Nijverheidsstraat 48c,
B-2260 Oevel,
Belgium
T: +32(0)14.46.21.10
F: +32(0)14.46.21.11
[email protected]
Personal contact:
Christian Machens
[email protected]
Figure 5-6. A Hydrogenics electrolyzer.
114
Table 5-4. Hydrogenics electrolyzers.
Hydrog
enics
Specifications
of 10 / 25
Bar
Electrolysers
(no
external
equipment,
gas
separator, cooling
pumps,
power
supply, etc.)
Max capacity
Nm3/h
Max power
kW
Kwh
/
Efficiency
Nm3 H2
Number of cells
DC Voltage
V
DC Current
Current density
Amp
A /cm2
Outlet Pressure
Bar
from x
% to y%
dA / dt
%
hydroge
n
Operation range
Operation speed
Purity
Weight
Diameter
Circulation
electrolyte
KOH
Lifetime
Operation
temperature
Price
nam
e
/mo
del
nam
e
/mo
del
100
0/
60
nam
e
/mo
del
100
0/
90
nam
e
/mo
del
100
0/
120
nam
e
/mo
del
400
0/
50
nam
e
/mo
del
400
0/
100
nam
e
/mo
del
400
0/
150
nam
e
/mo
del
400
0/
200
60 90 120 50 100 150 200
252 378 504 210 420 630 840
4.2 4.2 4.2 4.2 4.2 4.2
4.2
100 100 100 100 100 100 400 400 400 400
0
0
0
0
0
0
0
0
0
0
10
/25
25 100
10
/25
25 100
10
/25
25 100
10
/25
25 100
10
/25
25 100
10
/25
25 100
10
/25
25 100
10
/25
25 100
10
/25
25 100
10
/25
25 100
99.9 99.9 99.9 99.9 99.9 99.9 99.9
99.
9
99. 99.
9
9
260
0
Kg
m
of
Maintenance costs
Service time
Anode material
Cathode material
Diaphragm material
Price
Price
nam
e
/mo
del
100
0/
100 100 45
0/
0/
15 30
15 30 45
63 126 189
4.2
4.2 4.2
nam
e
/mo
del
l/min
%
Years
30
30
30
30
30
30
30
30
30
Celsius
USD/
year
Years
1000
Euro
USD
1000
DKK
531
1.05
5
3.98
3
7.91
3
331
2.4
83
115
30
5.1.1.3 IHT
Clos-Donroux
C. P. 228
1870 Monthey 1
Switzerland
T: +41 24 471 92 57
F: +41 24 471 92 64
[email protected]
www.iht.ch
Personal Contact:
Ernest Burkhalter
[email protected]
Figure 5-7. Large scale alkaline electrolyzer from IHT.
Figure 5-8. Alkaline electrolyzer from IHT.
116
Table 5-5. IHT electrolyzers (elevated pressure)
Specifications of Electrolysers Iht, 32 Bar
(no external equipment, gas
separator, cooling
pumps, power supply, etc.)
Max capacity
Nm3/h
Max power
Efficiency
Number of cells
DC Voltage
DC Current
Current density
Outlet Pressure
Operation range
Operation speed
Purity
Weight
Diameter
Circulation of electrolyte
KOH
Lifetime
Operation temperature
Maintenance costs
Service time
Anode material
Cathode material
Diaphragm material
Price
Price
Price
kW
Kwh / Nm3 H2
V
Amp
A /cm2
Bar
from x % to y%
dA / dt
% hydrogen
Kg
m
l/min
%
Years
Celsius
USD/ year
Years
name
/model
name
/model
name
/model
Lurgi
system
110 to 760
473
to
3268
4.3
32
25-100
99.9
Euro
USD
DKK
117
Table 5-6. IHT electrolyzers (abiment pressure).
Specifications of Electrolysers IHT 1 Bar
(no external equipment, gas
separator, cooling
pumps, power supply, etc.)
Max capacity
Nm3/h
Max power
Efficiency
Number of cells
DC Voltage
DC Current
Current density
Outlet Pressure
Operation range
Operation speed
Purity
Weight
Diameter
Circulation of electrolyte
KOH
Lifetime
Operation temperature
Maintenance costs
Service time
Anode material
Cathode material
Diaphragm material
Price
Price
Price
kW
Kwh / Nm3 H2
V
Amp
A /cm2
Bar
from x % to y%
dA / dt
% hydrogen
Kg
m
l/min
%
Years
Celsius
USD/ year
Years
name
/model
name
/model
name
/model
Bamag
system
3 to 330
11.7
to
1287
3.9
10 to 100
1
25 to 100
99.8
Euro
USD
DKK
118
5.1.1.4 AccaGen SA,
Via San Mamete,
CH-6805 Mezzovico
Switzerland
T +41 91 940 21 11,
F +41 91 940 21 04,
[email protected]
www.accagen.com
Personal contact:
R. Dall'Ara, CEO
[email protected]
Figure 5-9. Electrolyzer unit from AccaGen.
119
Table 5-7. AccaGen electrolyzers
Specifications
of
Electrolysers
(no external equipment, gas
separator, cooling
pumps, power supply, etc.)
Max capacity
Max power
Efficiency
Number of cells
DC Voltage
DC Current
Current density
Outlet Pressure
Operation range
Operation speed
Purity
Weight
Diameter
Circulation of electrolyte
KOH
Lifetime
Operation temperature
Maintenance costs
Service time
Anode material
Cathode material
Diaphragm material
Price
Price
Price
AccaGen name name name
6 / 10 / 30 /mod /mod /mod
el
el
el
Bar
AGE AGE AGE
1.0
2.5
5
name
/mod
el
AGE
10
Nm3/h
1
kW
4.8
Kwh
/ 4.8
Nm3 H2
10
45
4.5
2.5
11.8
4.7
5
23.2
4.6
name name name
/mod /mod /mod
el
el
el
AGE
20
20
89
AGE
50
50
222
AGE
100
100
440
4.5
4.5
4.4
V
Amp
A /cm2
6/10/ 6/10/ 6/10/ 6/10/ 6/10/ 6/10/ 6/10/
30
30
30
30
30
30
30
Bar
from x %
to y%
dA / dt
%
99.8
hydrogen
Kg
m
l/min
%
Years
Celsius
USD/
year
Years
99.8
99.8
99.8
99.8
99.8
Euro
USD
DKK
120
99.8
5.1.1.5 Erre Due
Agent:
H2Industrial ApS
Tjelevej 42
7400 Herning
Denmark
T +45 9627 5607
F +45 9714 0899
www.h2industrial..com
Personal contact:
Jesper Nissen Boisen
[email protected]
Figure 5-10. Erre Due electrolyzers. Here distributed by H2Industrial.
121
Table 5-8. Erre Due electrolyzers
Specifications of Electrolysers
(no external equipment, gas
separator, cooling
pumps, power supply, etc.)
Max capacity
Max power
Efficiency
Number of cells
DC Voltage
DC Current
Current density
Outlet Pressure
Operation range
Operation speed
Purity
Weight
Diameter
Circulation of electrolyte
KOH
Lifetime
Operation temperature
Maintenance costs
Service time
Anode material
Cathode material
Diaphragm material
Price
Price
Price
Erre Due / H2Industrial
name
/model
name
/model
Nm3/h
kW
Kwh / Nm3 H2
32.00
21.33
108
5.1
64.00
42.63
213
5.0
4
4
99.8
2700
99.8
1,000,000
1,450,000
V
Amp
A /cm2
Bar
from x % to y%
dA / dt
% hydrogen
Kg
m
l/min
%
Years
Celsius
USD/ year
Years
Euro
USD
DKK
name
/model
122
5.1.2 Eastern marked
During the search for electrolyser companies contacts to several scientists,
businessmen and also with Commercial Consulates Offices in Russia and Ukraine
were made.
The most interesting company Uralhimmash has provided the most detailed
information. The answering of the form was followed up by a visit to the factory in
Elaterinbourg, Ural, Rusia.
Similarly to the Western market of electrolysers, the market in East Europe seems to
be dominated by only few reputable companies - located in Russia.
Based on our business trip to Ekaterinburg it became likely that Uralkhimmash has a
monopoly position in the industrial alkaline electrolyzers production in Russia and
former USSR countries.
123
5.1.2.1 Uralhimmash
JSC “Uralkhimmash”
Khibinogorsky per. 33
620010 Ekaterinburg
Russia
Contact person:
Andrey Arkadyevich
Director of the direction of the electrolysers
T: (343) 221-61-55
E: [email protected]
www.uralhimmash.ru
Figure 5-11. Electrolyzer system from Uralkhimmash.
124
Table 5-9. Uralhimmash electrolyzers
name name name name
Uralhi
name name /model /model /model /model name
Specifications
of m/model
Electrolysers
mash /model /model
(no external equipment,
gas separator, cooling
SEU
SEU
SEU
FV
FV500
pumps, power supply,
etc.)
SEU 4
10
20
40
250M
M
BEU
125,
250
Max capacity
Nm3/h
4
10
20
40
250
500
625, 1
Max power
kW
20, 6
50
102, 5
205
1 520 3 150
250
Kwh /
Nm3
H2
5, 15
5
5
5
5, 3
5, 3
5
Efficiency
3х100,
Number of cells
30
25
50
100
82
166
6х100
DC Voltage
V
75
60
115
230
450
850
230
DC Current
Amp
330
1 000 1 000 1 000 8 000 8 000 1 000
A
Current density
/cm2
Outlet Pressure
Bar
10
10
10
10
1
1
10
Operation range
Operation speed
dA / dt
%
hydrog
en
99
99, 7
99, 7
99, 7
99, 5
99, 5
99, 7
Purity
7453х
3,
10136 7435х
6
Weight
Kg
1290
3390
4720
7435 59420
0
1,6х1, 1,6х1,
Diameter
m
0,46
0,89
0,89
0,89
9
9
0,89
Circulation of electrolyte l/min
300400
g/litre
KOH
%
Lifetime
Years
20
20
20
20
210
210
20
Operation temperature
Celsius
85
85
85
85
85
85
85
USD/
Maintenance costs
year
Service time
Years
6
6
6
6
6
6
6
Anode material
Fe
Fe
Fe
Fe
Fe
Fe
Fe
Cathode material
Ni
Ni
Ni
Ni
Ni
Ni
Ni
Diaphragm material
асбест асбест асбест асбест асбест асбест асбест
Price
Euro
Price
USD
Price
DKK
125
5.1.3 Prices, Efficiency, CE-Marking, Safety
Prices
Very large plants have been installed for production of fertilizer in countries with
cheap hydropower, op to more than 100 MW in capacity. A third common application
in Eastern Europe and Russia is the use of hydrogen for cooling of power plant
generators from on site electrolysers.
Though the technology is well known and mature the price is too high for
energy applications. The reason is the market is limited and there are just a few
suppliers to cover the world marked. A rough estimate is a magnitude of 10 MW pr.
years.
When in the feature electrolysers will be used in grids with a large amount of
wind power 10 MW at least, will be the magnitude of a single plant necessary to
balance just one wind farm.
Today the marked is characterized by few plants sold annually at a high prices
for industrial use, contrary to what we will see in the future where a large number of
plants for energy use will be sold at a very low price.
If we look at the development of prices of wind turbines we will get a good picture of
what will happened when energy plants is produced in large numbers. During the past
20 years the prices of wind power is reduced to 20% of what it was 20 years ago.
There is no reason not to believe that we will see the same development for
electrolyser plants.
Electrolysers have the reputation of being very expensive. It is true but often
when the price pr. kW of a specific electrolyser is mentioned the size of the plant is
not given. The specific price of electrolysers (EURO / kW) is strongly dependent of
the size of the plant.
The price analysis below shows it very clearly. It can be seen that the price per
kW installed capacity vary with a factor of 10 dependent of the size of the plant.
126
Prices of Alcaline Electrolysers
10000
9000
8000
Euro / kW
7000
6000
5000
4000
3000
2000
1000
0
0
500
1.000
1.500
2.000
2.500
3.000
kW
1 Bar, HYDRO
4 Bar, H2Industrial
16 Bar, HYDRO
6 Bar, AccaGen
10 Bar, Hydrogenics
1 Bar ELT
Figure 5-12. Prices of electrolyzers as a function of power rating.
The prices are collected over the last 5 years from year 2000 and therefore not
consistent. Electrolyzer plants are often tailor-made and directly price comparison
between the different manufactures is therefore not possible. Anyhow, the graphic
shows clearly that in order to obtain relatively cheap electrolyzers, they have to be as
large as possible and not smaller then about 1 MW.
The company H2industrial (Erre Due) seems to disturb the picture, but
because they only make small plants and with a low efficiency (69%) it is not so
important for this study.
Efficiency
The efficiency of an electrolyser is defined as the ratio of the higher heat value (HHV)
of the hydrogen produced and the DC electricity consumption of the electrolyser.
Simple electrodes made of mild steel and coated with nickel have an
efficiency of about 68% and the most advanced experimental electrodes manufactured
by vacuum plasma spray technique have reached efficiencies as good as 90%.
The commercial electrolyzers have an electricity consumption of 4.1 to 4.8
kWh per normal m3 produced. Using the HHV of 3.5 kWh/m3 hydrogen, the
efficiencies can be calculated to between 85 and 73%.
Below is showed the efficiency of the electrodes from Iht, Norsk Hydro,
Hydrogenics, Uralhimmash and the VPS electrodes from DLR, Stuttgart, Germany.
The electrolyser from Aarhus University, HIH is also shoved. It is a lab model
with stainless steel electrodes, which will be converted with Vacuum Plasma Spray
electrodes from DLR.
127
HIH
Uralhimmash
Iht.
Hydrogenics
Norsk Hydro
Figure 5-13. A plot of efficiency of the electrodes from Iht, Norsk Hydro,
Hydrogenics, Uralhimmash and the VPS electrodes from DLR, Stuttgart, Germany.
The HHV is always used when calculating the efficiency of electrolysers whereas the
lover heat value (LHV) of 2.9 kWh/m3 is used when the efficiency of fuel cells is
calculated. The reason is that the fuel cell is consuming hydrogen and therefore the
calculated numerically efficiency will be higher when the LHV is used.
Safety
Safety is an important issue for electrolyser plants and will be handled by the CEmarking. Especially two directives will secure the necessary safety level regarding the
hydrogen as a potential explosive pressurized gas. It is:
1) DIRECTIVE 94/9/EC concerning equipment and protective systems intended
for use in potentially explosive atmospheres.
2) DIRECTIVE 97/23/EC concerning pressure equipment
Two other directives will be involved in the approval of the power supply and the
control system. That is:
128
1) Council Directive 73/23/EEC relating to electrical equipment designed for use
within certain voltage limits.
2) Council Directive 89/336/EEC relating to electromagnetic compatibility
5.2 Cost
5.2.1 Hydrogen as an energy raw material
The question is not how many years the remaining oil and gas will last, but when will
the supply not be able to meet the demand.
According to the European Hydrogen Association (se graphics below) this is
going to happen within the next 10 to 20 years. But before then, there will be a need
for alternative fuels for domestic heat and power as well as for transport. Due to the
shortage of supply the prices of fossil fuels will go up and the alternatives will
become competitive.
Figure 5-14.
Compressed hydrogen distributed via pipelines can be used for supply of fuel cells
integrated in CHP plants (Combined Heat and Power).
Plant oil and synthetic fuels made from biomass is an evident option for the transport
sector, however, the resources will not be sufficient to meet the demand.
129
Hydrogen as a compressed gas can be used in special designed vehicles as a
supplement to ethanol used in conventional cars, but there are also other possibilities.
Hydrogen can be used as a new energy raw material in the production of different
synthetic fuels such as methanol, methane and ammonia.
In the fermentation process of biomass to ethanol one third of the carbon is
lost in the form of CO2. Since carbon from biomass will be short in supply this source
of carbon as CO2 must be utilized, which is possible by reacting with hydrogen.
Methanol or methane can be produced in this way.
Ammonia can be used as a fuel, but it is also used for NOx reduction on coal
and biomass-fired power plants.
Figure 5-15. Hydrogen from water electrolyses used for production of synthetic fuels.
5.2.2 Electricity costs
Hydrogen is produced from water and electricity. The water consumption is about one
litre per normal cubic metre (Nm3) of hydrogen and the electricity required is approx
4kWh per Nm3. This means that the water price is minimal compared to the price of
the electricity. This is true even though the water needs to be purified in order to
remove traces of salts and organic residues that will otherwise accumulate in the
electrolyzer.
Electricity is traded at hour-to-hour prices on the spot market and the prices
vary from 0 to 15 Euro Cent per kWh. It is therefore obvious to use electricity for
hydrogen production during the cheap hours. On Figure 5-16 the prices of electricity
at the West Danish spot market is showed for every hour in 2006.
130
Figure 5-16. The price of electricity at the West Danish spot market hourly through
2006.
If all hourly prices are put in order starting with the lowest price and the highest price
last, we get a very clear impression of the price frequency. From Figure 5-17 it can be
seen that there is only few hours with prices below DKK 0,20/ kWh and above DKK
0,50/ kWh. 83% of the time the prices are between these two values.
131
Figure 5-17. The price of electricity at the West Danish spot market through 2006
arranged.
Figure 5-18 shows the average price of the electricity if an electrolyser is operated a
certain number of hours in 2006, and the operator has purchased the cheapest
electricity on the market. It can be seen that when the electrolyser is operated nonstop throughout the year the average price was DKK 0,33/kWh. If the electrolyser
was in operation 50% of the time, corresponding to 4380 hours, then the price was
DKK 0,25/ kWh.
A strategy to obtain the cheapest electricity could be to buy in the off-peak
period from 9 p.m. to 6 a.m. The number of operating hours will then be 8760 x
(9/24) = 3285 hours and the average price would be DKK 0,23.
132
Figure 5-18. The average price of electricity at the West Danish spot market if an
electrolyzer is operated a varied number of hours through 2006.
5.2.3 Investment costs
From the Figure 5-19, it can be seen that the relative price of electrolysers is strongly
dependent on the size. The larger the plants are the smaller prices per kW installed.
Electrolysers are manufactured in modules up to 1 to 3MW. Therefore the price curve
becomes almost flat after this size. For smaller plants price reduction is obtained by
up-scaling the electrolyser stack instead of using more modules. Therefore, the price
elasticity is much higher for the smaller plants.
133
Euro / kW
Prices of Alcaline Electrolysers
10000
9000
8000
7000
6000
5000
4000
3000
2000
1000
0
0
500
1.000
1.500
2.000
2.500
3.000
kW
1 Bar, HYDRO
4 Bar, H2Industrial
16 Bar, HYDRO
6 Bar, AccaGen
10 Bar, Hydrogenics
1 Bar ELT
Figure 5-19. Prices of the different electrolyzers reviewed in chapter 5 as a function
of production rate capability.
5.2.4 Depreciation
The investment costs for an electrolyser plant in MW size is between EUR 500 and
1500 per kW. This price, say 700 Euro/kW for a 2 MW plant must be depreciated
over the hours of operation. Therefore, the more hours the plant is in operation per
year the less depreciation per hour of operation. This is due to the fact that there are a
maximum number of years for the depreciation of the plant. For example 10 years.
As an example the depreciation of the largest electrolyser from Figure 5-19 is showed
in Figure 5-20 as a function of number of operating hours per year and a lifetime of
10 years.
134
Figure 5-20. Cost of electrolyzer pr kWh produced assuming a 10 year depreciation.
The calculation is based on the largest electrolyzer from Figure 5-19 and as a function
of operating hours.
5.2.5 Optimum operation of electrolyser plants
The price of the hydrogen produced is mainly calculated from the sum of the price of
electricity and the depreciation. Since the average price of electricity increases by the
number of hours of operation per year and the depreciation decreases by the number
of operating hours, there will be an optimum number of operating hours per year.
Figure 5-21 shows that the lowest hydrogen price will be obtained if the plant is
operated approx 50% of the hours through out the year. These hours are most likely to
be during the night, because the lower consumption from the industry and the private
homes will cause the electricity price on the spot market to go down. However, the
curve is quite flat at operating time up to 100%, so if it serves a purpose to increase
the production the additional cost is only marginal. This might well be desired when
the demand for fuel is growing in the transport sector.
135
Figure 5-21. The cost of hydrogen produced by electrolysis based on the assumptions
above. Key figures are listed in Table 5-10.
All applied input data for the analysis is showed in Table 5-10.
Table 5-10. Input data for the analysis of production cost.
Installed capacity MW
Size of plant Nm3/h:
Price of plant, DKK.:
Efficiency, kwh/Nm3 DC:
Burning value for hydrogen kWh/Nm3
Installed capacity MW
Years of operation:
Operation costs, DKK/kWh:
Maintenance costs, DKK/kWh:
Efficiency of hydrogen production
2877
620
12,000,000
4.64
3.5
2877
10
0.01
0.01
0.75
5.2.6 Utilization of oxygen and heat.
The electrolyser also produces heat and oxygen. 1m3 of oxygen is made for each 2m3
of hydrogen. In this example the hydrogen is produced at an efficiency of 3.5/4.64 =
0.75. The remaining 25% is converted to heat. If it is assumed that 90% of this heat
can be utilized for district heating then the economy for utilizing oxygen and excess
heat from the electrolyser can be calculated as shown in Figure 5-22.
136
Figure 5-22. The cost of hydrogen produced by electrolysis based on the assumptions
above taking the value of produced oxygen and heat into account. Key figures are
listed in Table 5-10 And Table 5-11.
The applied data is showed in Table 5-11.
Table 5-11. Key figures for calculation of the value of the produced oxygen and heat.
Heat production MW/MW
Price of heat DKK/MWh
Sale of heat DKK/MWh electricity consumption
Oxygen production Nm3/MWh electricity consumption
Price of oxygen DKK/m3
Sale of oxygen DKK/MWh
0,22
236
52
108
0,5
54
Assuming 100% electrical efficiency
The influence on the price of hydrogen, if the efficiency of the electrolyser is
increased from the quite low 75% used in the example to the maximum possible
100% possible in the future, is shown in Figure 5-23. Only the efficiency of the
electrolyser is changed, all other input data are the same as in previous calculations
above. This shows that the price of hydrogen per kWh is reduced by 10% when the
efficiency of the electrolyser is raised from 75% to 100%.
This is not an argument against research and development of more efficient
electrolysers, but a very strong indication that there is absolutely no reason to await
more efficient electrolysers to start business development in this very promising
energy technology.
137
Figure 5-23. Calculated cost of hydrogen like in Figure 5-23 with the only difference
that 100% electrical efficiency is assumed.
5.2.7 Excess of wind power
It is often claimed that excess of wind power can be used for low-cost hydrogen
production. This is not correct. There is no excess of power in the grid. All the
electricity produced is consumed, but at different prices. This analysis shows that the
numbers of hours with very low prices are quite few and far from enough to give a
reasonable depreciation of the plant and price of the hydrogen produced.
5.2.8 An estimate of the European market for electrolysers
A European market for electrolysers for demonstration projects has already been
established. According to EU official figures, the Commission and the national and
regional funds have invested EUR 160 million a year in hydrogen related research ,
development and demonstration in the period from 2002 to 2006. This figure is
corresponding to the statements from the director of the Energy Research, European
Commission, Pablo Fernandes-Ruiz, who expects the figure to be doubled in 2007.
138
Figure 5-24. The distribution of the European investments. From “Energy Research
European Commission”.
From Figure 5-24 it can be seen that 15% is spent on hydrogen production. If we
estimate that one third of the expense is used on electrolyser systems the potential
market can be calculated as: 160 x 2 x 1/3 x 0,15 million Euro = 16 million Euro
annually equivalent to DKK 120 million.
5.3 Visits to suppliers of electrolysers
When asking manufacturers of electrolysers for information, we quickly realised that
asking questions without having a project was a very bad idea. The electrolyser
companies are relatively small, and they do not have the resources to answer
questions from students, research institutions and others looking into a coming
hydrogen society in a proper way. For this reason we decided to focus on electrolysers
to fit into production of synthetic fuel.
When looking for commercially available electrolysers > 100Nm3 H2/hour the
only choice is the alkaline electrolyser.
The idea was to contact and visit some of the leading commercial suppliers of
electrolyser equipment, to examine units in operation, to look into reference lists, to
talk with maintenance people, and to get our own impression of what an electrolyser
is.
The overall agenda for the meetings was:
We are an Electricity Utility Company in Denmark looking into solutions on
producing synthetic fuel for the transport sector based on biomass and wind power.
We are contacting you about the electrolysis part to get more knowledge about
the possibilities now and the possibilities in future.
139
•
•
•
Efficiency
Pressure
Technology
For a pilot plant we are looking for an amount of 125Nm3 H2/hour. For a full-scale
project we need 50000Nm3 H2/hour.
The H2 will be a component for the methanol production. This process will run
at an estimated pressure of minimum 70 bar, which means that a high-pressure H2
production will be an advantage.
We hope to talk to you to discuss the actual needs and to get information about
what is going on in your electrolyser business.
We visited the following companies:
•
•
•
•
Norsk Hydro, Nottodden, Norway
IHT (former Lurgi), Geneva, Switzerland
Acca-gen, Lugarno, Switzerland
Hydrogenics, Antwerp, Belgium
The detailed agenda for the visits was according to the questions and minutes of
meeting below.
5.3.1 General questions and answers.
The electrolyser is foreseen to stabilise the electrical grid and make it possible to
connect more wind power to the grid. For this reason, the electrolyser duty will be in
the area of 4000-5000 full load hours a year and the electrolyser will often be
disconnected more times daily for short or long periods; up to 2000 hours a year.
•
Will that be an acceptable way of running an electrolyser?
o As long as the electrolyser is kept hot, and the load gradients are inside the
limits, it will be acceptable to the electrolyser. For pressurised electrolysers it
is advised to depressurise during hours with no load.
•
Will this way of operating the electrolysers generate more maintenance?
o Some suppliers have experience from sun- and wind-powered systems and
have developed special electrodes able to sustain sudden and frequently
repeated current interruptions without the need for polarization maintaining
current during long shutdowns.
Our plan is to start production of synthetic fuel. We expect to start up with a pilot
production to get more experience. For this purpose we need an amount of 125Nm3
H2/hour, (about 625kW). For the full-scale production we need 40000-80000Nm3
H2/hour (approx 180-360MW).
•
What is the optimum size of an electrolyser for that purpose?
o The electrolyser capacity on the market today is mainly below 200Nm3
H2/hour. Even though units close to 1000 Nm3 H2/hour are already available.
140
However, this limit is not a technological one but is commercially imposed by
the competition from stem reforming which is by far more economic. In the
past electrolysers with capacities up to 20,000Nm3/hour have been built
(100MW electrical absorption). In the future, if the demand for large
quantities of clean hydrogen increases, larger electrolyser units could be built
and in this way the cost would be strongly reduced. Developing electrolysers
with capacities in excess of 1000Nm3/hour is just an engineering matter and
not a feasibility question. A reasonable size could be 5000Nm3/hour.
Figure 5-25. The parallel and size law.
The H2 will be the main component for the synthetic fuel production; this process is
running at an estimated pressure of minimum 70 bar.
•
What could be the optimum pressure for the electrolyser?
o If the electrolysis process is carried out in a closed environment, the
gas produced by water splitting will increase the pressure. For small
electrolysers, the feasibility of very high-pressure electrolysers has
been shown commercially. In fact devices up to 200bar have been
built. From experience, however, it has been demonstrated that very
high-pressure electrolysers must be limited in size in order to be price
competitive compared with solutions with a low-pressure electrolyser
and a gas compressor.
141
Figure 5-26. Suitable electrolyser pressure as a function of sizes.
Running an electrolyser the consumption of electricity is a very important economical
factor, for that reason the efficiency of the electrolyser is a very important point.
•
What is the efficiency of the alkaline electrolyser?
o Standard electrolysers consume about 4.5 kWh/Nm3 of H2 at 100% load,
but the consumption must always be compared to the current density. This
value can be further improved down to 4.0 kWh/Nm3 at the expense of an
increased electrolyser cost or by running at part load.
During periods with strong wind and low consumption in the grid, cheap electricity
may be available.
•
Will it be possible to run the electrolyser at overload during these
periods?
o If the feed water and cooling capacities are available, 20% overload
will normally not be a problem, but at a lower efficiency.
The electrolyser efficiency is a split between H2 production and the cooling effect. If
the process temperature in the electrolyser is raised, the cooling water could be
utilised for district heating or for evaporation of water in the ethanol process.
•
What is the normal process temperature for a catalytic electrolyser?
o The normal operating temperature for commercial alkaline electrolysers is
in the range from 70ºC to 90ºC.
•
Will it be possible to raise the temperature?
o It would be convenient to raise the temperature because it will increase the
efficiency of the electrolyser. The limiting factor is the corrosion. For a
traditional alkaline electrolyser it is not possible to raise the temperature
without a dramatic change of design and material selection.
142
The electrolyser is foreseen to stabilise the electrical grid and absorb the excess of
wind energy. The electrolyser must be able to follow the fluctuation from wind power
which may be very rapid. What is the range of operation?
•
How is the reliability/maintenance for an electrolyser?
o Electrolysers are generally designed for more then 20 years of operation.
Only a short stop a year for maintenance is recommended. Some
manufacturers recommend a large overhaul every five to seven years.
There are units which have been in operation for more than 20 years
without the cells having been opened. Overhaul can be done on site, or the
stack may be sent back to the factory.
•
Does the electrolyser contain asbestos?
o In the past asbestos has been widely used as diaphragm in the electrolyser.
Today it is mostly phased out and substituted by other materials, but it is
still used by some manufacturers.
•
Which metals are contained in the electrolyser?
o The alkaline electrolyser is mainly made of carbon steel, nickel-plate
carbon steel and stainless steel, and it does not contain precious metals.
•
Will the electrolyser be able to run at part load?
o The electrolyser is basically able to run a load from 0 to 100%. At which
load is the electrolyser able to produce the guaranteed purity of gas? The
manufacturers state a range between 5-100% and 25-100% load,
dependent of the design regarding to stray currents.
•
How quickly is it possible to change the load of the electrolyser?
o If the electrolyser is on operation temperature and pressurised, the
manufacturers state a load ramping time (0% to 100% load) from 20 sec
to 10 minutes. The challenge seems mainly to be to maintain constant
thermal conditions in the electrolyser. Quick ramping demands advanced
temperature control of the cooling circuit.
•
Which codes and standards are followed for the production and approval
of the electrolyser?
o The manufacturers comply with relevant EU directives such as ATEX
(94/9/CE) and PED (97/23 CE).
•
ATEX: What will the classified area look like for an electrolyser?
o The hazardous zone may be dependent on the ventilation, which can be
either natural or forced. For large installations we have only seen natural
ventilation, where the top of the roof has been open for output. Forced
ventilation has been chosen for turnkey container installations. For the
room containing the electrolyser and small extensions outside the building,
the classification will, as a minimum be zone 2 IIC T1 according to EN
60079-10. The explosion-proof components will be selected as EEX d IIC
T1 or EEX ia IIC T1.
143
•
What kind of weather protection is needed for the electrolyser?
o Only light buildings are needed to protect against frost. When the
electrolyser is in operation, no external heating is needed.
•
Will control and safety be independent for the electrolyser?
o As standard, the electrolyser will be equipped with a PLC system for
monitoring and control. For safety functions the electrolyser will be
equipped with an independent, hardwired or SIL-classified safety system
able to execute a tripping of the unit. The selection of safety system will be
based on a risk assessment.
•
How is the load of the electrolyser controlled?
o The load of the electrolyser is controlled by the current which is
proportional to the load.
144
5.3.2 Norsk Hydro minutes of meeting.
Time, place: 8-9 June 2005, Rjukan and Notodden
Participants:
Elsam:
NH, KHI, NTO
Norsk Hydro: Andres Cloumann (AC, Sales Manager Hydro Electrolysers), Dag
Øvrebo (DØ, Development Manager, Hydro Technology Ventures),
Knut Harg (KH, Managing Director, Hydro Electrolysers)
Visit to production plant in Rjukan
At a hydrogen peroxide factory in Rjukan, Akzo Nobel has erected four (atmospheric)
electrolysis modules (of slightly different technological generations. Three of the
modules have a capacity of 500Nm3 H2/hour, one module has a capacity of
300Nm3/hour. The following information was given: The length of the large modules
is 11-12 metres, and they have a diameter of approx 2 metres. The modules were
mounted on skids. Max 230 cells per frame/skid. The electrolysers perform approx
6,500 operating hours a year.
In general, overhaul of the cell stacks is performed every seven years. During
an overhaul, the stack is separated on the spot and electrodes/membranes and worn
plates are replaced with reconditioned plates. In connection with an overhaul, each
stack is taken out of operation for approx one week.
The routine maintenance requirement of the electrolyser is minimal.
“Standard” maintenance is required for lye circulating pumps, etc. It is not at all
necessary to perform daily service inspections – remote monitoring of the systems is
possible.
Feed water is supplied from a RO plant. In the plant in question, the gas
pressure was supplied by a water ring compressor, which also purified the gas of the
remaining lye, which was subsequently re-circulated. The requirement for addition of
“new” KOH is very small, but the water quality is very important – especially the
carbonate content – to the speed with which impurities are formed in the lye cycle –
and the consequential efficiency reduction.
Gas quality is checked by measuring the oxygen content in the H2 flue gas and the
hydrogen content in the O2 flue gas.
O2 in H2 below 0.2% (during the visit, the concentrations recorded on the operating
plant were 0.11-0.14%)
H2 in O2 below 0.5% (during the visit, the concentration recorded on the operating
plant was 0.443 %).
The guaranteed control range of the plant is 20-100% load. The maximum and
minimum functions of the electrolyser have not been fully tested, because in general it
is not necessary to operate the plant to the limits absolute max/overload and minimum
load. The gas production follows almost immediately once the power is connected,
but there is a lag time in the system before the gas pressure starts building up and gas
production is initiated. This is mitigated by inserting a gas bell for compensation
purposes. During brief production outages the gas pressure in the electrolyser may be
maintained to ensure that the gas production starts as soon as the power is connected.
145
The electrolyser can be started and produce in the load range almost instantaneously
provided that the temperature of the electrolyte is maintained at approx 80°C.
The optimum operating temperature is 80°C (caustic breaking, generally in
pipes and pumps, not in the electrolyser). The temperature is controlled by cooling the
lye.
Gas analysis outlet branches are installed for every 20 cells – may be used for
troubleshooting in case of reduced gas quality.
ATEX. The production hall with electrolyser is classified as zone 2 gas group IIC T1.
The production hall has natural air ventilation. Air inlets are installed along the walls
and outlet is installed in the roof ridge in the entire length of the building. When the
electrolysers are being kept warm or when they are operating, the heat generated will
promote the natural ventilation.
Visit at Norsk Hydro in Notodden, where the electrolysers are assembled and
reconditioned. Norsk Hydro’s development of electrolysers is also located here.
The electrodes are made of a nickel plate steel supporting plate. On one side, the
cathode is mounted, on the other side of the supporting plate, the anode is mounted.
Cathode and anode are both perforated steel plates with an (unknown) surface
coating. The surface coating contributes to the high efficiency.
Membrane. Separates anode and cathode. Must be able to stand the aggressive
chemical environment, be permeable to ions/electrons and have a certain minimum
gas tightness. Is typically made of asbestos – Norsk Hydro, however, makes their
membrane of a different material (unknown). It looks like woven cloth – maybe a
type of flour polymer.
Efficiency. Comment from Andres: “4.1kWh/Nm3 can be accomplished by everyone.
You must consider the current density to assess the actual efficiency. What is more
relevant though is that the electrolyser is able to increase production by 25% at an
additional consumption of 0.25kWh/Nm3.
Electrolysers are generally supplied as turn key supplies including transformer,
electric switchboards, C&I and safety system.
Pressurised electrolysers: Hydro is currently developing a pressurised electrolyser (30
bar). The test plant is expected to be completed by end 2005, and is subsequently
ready to be long-term tested at a customer, e.g. in a pilot plant for production of
synfuel at Elsam (Proposal: Hydro owns the plant, we operate it, pay for the
electricity and own the gasses produced. Data is divided). The design of the test plant
was very compact, and furthermore, the response times on the gas side were good due
to the low dead volume in pipes and pressure tank. The control range is 10-100%.
Hydro expects future plants for large-scale production to be pressurised. The
construction costs of a pressurised electrolyser will be higher – in each case, the
decision between pressurised/atmospheric electrolysis will depend on the price of the
compressor.
146
It takes 8-10 months from the order is placed to manufacture a demonstration plant
(~500kW).
5.3.3 AccaGen, minutes of meeting
Time, place: 29 September 2005, Via San Mamete (CH)
Participants:
Elsam:
KHI, NTO
AccaGen:
Damian Gamma (Sales and Marketing), Roberto Dall`Ara
(Director)
Three years ago, the company AccaGen was divested from the company Casale,
which among other things designs/supplies NH3 factories. The divestment was a step
towards enhancing focus on the electrolysis business, which compared to Casale’s
other business activities only constituted a very small part. However, AccaGen is still
able to benefit from Casale’s resources and know-how.
The company produces alkaline pressurised electrolysers, generally in sizes from 0.1100Nm3/h H2.
The company employs 10 people. Production of components is handled by
sub-suppliers. Assembly and testing is handled at the factory.
The company participates in a number of research activities in cooperation with
various universities.
AccaGen only manufactures pressurised electrolysers for minor plants up to
200 bar. For large plants > 50Nm3/h H2, the maximum pressure will be approx 30
bar. In large plants, the compressor price constitutes a relatively smaller part of the
plant costs, which is why large pressurised plants are less interesting.
The company does not think that the production of units with a capacity of up
to 5000Nm3/h H2 will constitute technical problems, and the company sees a market
for plants of this size. However, it requires more plants of this size for the market to
have a reasonable volume.
In relation to price, the following key figures are used in connection with
expansion of capacity. If the capacity is increased by a factor 10 by increasing the cell
size, the plant price will increase by a factor 6. The same expansion of units
connected in parallel will increase the plant price by a factor 9.
The typical industrial electrolyser will have a power consumption of 4.64.8kWh/Nm3 H2. In applications where it is deemed expedient, the efficiency may be
increased to 4.0kWh/Nm3 H2.
AccaGen has supplied electrolysers for stand-alone wind turbines and solar cell
installations, and has used these unconventional operating situations, which occur
when the sun hides behind a cloud or when the wind slows down for a period of time,
in the design of its control system. These are operating conditions which do not
normally occur in a plant operating at continuous load.
AccaGen were very eager to emphasise its DCS, because the use of electronic
solutions makes it possible to build more compact plants, as a number of balancing
tanks can be left out. The design philosophy was: ”Compact and easy-to-use”.
147
AccaGen observes the relevant European directives and standards, e.g. the
tank directive and ATEX, etc.
In case of co-production of O2 the plant shall be prepared for this from the
beginning. The O2 pressure will typically be 5 mbar below the H2 pressure.
AccaGen does not use asbestos products (for membranes). AccaGen’s
membrane is sturdy, but according to AccaGen it may be slightly less effective than
Hydro’s membrane.
AccaGen prescribes overhauls to be performed every 3-5 years. The design of
the electrolyser in which the cell rings are directly pressurised facilitates service and
maintenance to a certain degree. IPR may be a problem area which might prevent us
from performing large overhauls ourselves.
Operating range 5-100% without problems. AccaGen has developed an electrode
coating which is resistant to extensive standstill. This requirement emerged in
connection with supply for stand-alone wind turbines/solar cell installations. Frequent
on/off operation results in corrosion problems because the potential on the electrodes
change.
Around 50% of the production price of the total plant may be attributed to the
electrodes.
The normal temperature of the electrolyser is 80°C. AccaGen has no
experience operating the electrolysers at higher temperatures, but AccaGen is willing
to look into the possibilities. It was emphasised that the expected operating life of the
plants is at least 30 years at 80°C (provides a 20-year guarantee with a possibility for
a further 20-year extension).
Supplementary water must have a conductivity of less than 5μS.
Delivery time: 6-12 months.
Electrolysers are generally supplied as turn key supplies including transformer,
electric switchboards, C&I and safety system. They may also be supplied in a
container. The plant is designed for automatic operation, 8100 h/year for the isolated
systems.
148
5.3.4 IHT minutes of meeting.
Time, place:
Participants:
IHT:
HIRC:
DONG Energy:
30 October 2006 Geneva/ Monthey (CH)
Ernest Burkhalter, [email protected]
Hansruedi Arnold, [email protected]
Riland Helfer, [email protected]
Armando Hermosilla [email protected]
Lars Yde
Niels Tophøj
IHT produces, maintains and sells atmospheric electrolysers (Bamag) from 3 to
330Nm3/h and pressurised electrolysers (Lurgi) at 32 bat from 110 to 760Nm3. The
latter are the largest units in the market and at the same time they have the highest
discharge pressure.
The plants were developed in the 1950s by Professor Zdanski and no
fundamental changes have been made since then, except for the control system, the
safety system and the power supply. All plants are produced in the same place (ClosDonroux CP 228, CH 1870 Monthey 1); but under different names and owners:
Lanzola, Lurgi, Giovandola, GTec and now IHT. Ernest Burkhalter is director and
member of the board.
Technical questions and answers
Dynamic operation of the pressurised electrolysers, which are of a type that may
become interesting in connection with large electrolysis installations, is possible. It
takes about 10 minutes for a heated de-pressurised electrolyser to reach operating
pressure, after which the electrolyser may in principle be controlled freely between 0
and 100%. In practice, however, the requirement for a constant temperature in the
electrolyser means that a ramp is inserted. By optimising the cooling circuit, the ramp
time may be reduced to approx 1-2 minutes. The load range of the electrolyser is
stated as 25-100%, at lower loads problems may arise in connection with the purity of
the gas in relation to the data stated. The electrolyser can be shut down for up to 4-6
hours without losing pressure or temperature and without causing reduced operating
life. As long as pressure and temperature are maintained, load may be imposed on the
electrolyser.
At a power density of 200mA/cm2 and a cell voltage of 1.9V, the plant
consumes 4.61kWh/Nm3 H2 at full load. At part load, the consumption for each Nm3
H2 produced is reduced and the efficiency is increased. This means that it is better to
reduce the load on all electrolysers instead of shutting a part of the electrolysers
down.
The load curve was not stated.
The plant’s power draw on the power grid and thus its hydrogen production is
controlled via a PC-based DCS. The control is divided into a regular PLC part, where
the operation, trend curves, alarm management, etc of the plant are handled, and an
independent SIL (Safety Integrity Level) classified safety management system.
Power control is carried out as current control by means of controlled rectifiers that
rectifies the grid AC to DC and thereby supplying the electrolysers.
149
Gas qualities were stated as:
Hydrogen: 99.8-99.9%
Oxygen: 99.3 n-99.6%
O2 in H2: 0.1-0.2% vol.
H2 in O2: 0.4-0.7% vol.
KOH in the hydrogen gas: max 0.1mg KOH/Nm3:
H2O in the hydrogen gas: 1-2mg/Nm3
The gasses are free of: CO, CO2, CH4, S, Cl.
Feed water: 0.85 litre/Nm3
Cooling water: 40 litre/Nm3 at a temperature difference of 20°C.
The operating temperature is generally 84°C, but as a test, a non-asbestos electrolyser
has operated at a temperature of approx 95°C for a short period of time. IHT has not
tried to raise the temperature even further, but they were willing to consider the
possibility.
The diaphragms of the plant are made of 5mm asbestos. They are working on
using an alternative material based on nickel oxide. With this diaphragm, it is possible
to obtain a 12%-increase in the gas production with the same power consumption.
This is due to the fact that the diaphragm is thinner and thus the distance between the
electrodes is reduced. However, the impact on the operating life is unclear. The
lifetime of asbestos is more than 25 years.
The electrolysers mainly consist of nickel plate soft steel. There is possibly a
bit of molybdenum or other metal on the surface of the cathodes. Metals, which may
be in short supply in future, are as such not used.
The diameter is 1.6m. 139 electrodes are stacked into one unit, and the largest
electrolysers with an output of 760Nm3/h and a consumption of 3.5MW consist of
four units. One unit weighs approx 15 tonnes.
IHT are not planning on developing larger plants, because all parts for the
electrolyser are manufactured on special-purpose machinery, which is only able to
handle electrodes etc with a diameter up to 1.6m.
Electrodes and the bipolar plates have a service life of 15 to 20 years, after
which the plant is disassembled and the nickel is renewed. The service life of the
asbestos diaphragm is unknown. If the electrolyser requires an overhaul, it is shipped
unit-wise (139 cells) to the factory in Geneva. An overhaul is expected to take 2-3
weeks plus time for transport.
The plants may be supplied as turnkey projects in accordance with applicable
national standards. The plant is distributed in three separate rooms; one for the
electrolyser, one for the power supply, and one room for the DCS and safety systems.
The electrolyser room is classified as Eex zone 2.
The instrumentation was designed as an intrinsically safe installation, pump
engines were designed with increased safety and valve actuators were designed with
pneumatic actuators.
150
The building was designed for natural air ventilation with ventilation outlets at the
highest point of the building along the entire length of the building thus eliminating
the risk of local gas pockets.
Figure 5-27. Electrolyzer building with roof ventilation.
The operating voltage is ±540V on a 3.5MW unit. There is a common plus in the
centre and minus at each end. This means that electrically speaking the electrolyser is
made of two units connected in parallel. Each unit thus consume 3240A.
The plant at DJEVA in Monthey had a water-cooled transformer. It was a
requirement from the local utility company that filters were installed for the fifth,
seventh and the eleventh harmonics.
Figure 5-28. Filters.
151
The mechanical design of the stack
The electrodes are made as a woven net of approx 1.5mm thick threads. The
horizontal threads are completely straight while the vertical threads are woven around
the horizontal threads. This design makes the grid very compact, because the vertical
threads may be placed so close to each other that they touch. In this way a large
surface area of the electrodes is obtained. The anode, i.e. the oxygen electrode is a
nickel plate with 60µm. The cathode (the hydrogen electrode) also has an
electrochemical coating of nickel and probably zinc. Subsequently, the zinc is
removed with alkali, after which the cathode is activated. This results in a steep
increase of the specific surface area.
The electrodes are mounted very close to the diaphragm in a so-called zero
gab structure. The gas produced is led to the back of the electrodes and does thus not
reduce the conductivity between the electrodes.
An egg carton-shaped separator keeps the electrodes connected to the
diaphragm and at the same time, it functions as cell separator and provides room for
the gasses to rise and be led into the gas ducts at the top of the cell stack.
Separators and diaphragms are installed in steel rings which can subsequently
be stacked with seals in between. The electrodes are placed loosely between the
separators and the diaphragms. 139 cells are combined to one unit, and four units are
combined between two end plates with a thickness of approx 200mm to one
electrolyser stack.
The power density of the electrolyser may be calculated as max power
consumption divided by the volume of the electrolyser stack. With a diameter of
1.6m, a length of 10m and a power output of 3.5MW this is: 170kW/m3.
Wages, materials and capacity
It was stated that approx 50% of the costs are costs of materials. This is due to the
large increase in the prices of iron and nickel in recent years.
The production capacity is approx 3.5MW/plant/month corresponding to
approx 40MW annually.
152
Figure 5-29. Electrolysers from IHT of 3MW each.
153
5.3.5 Hydrogenics minutes of meeting.
Time, place:
Participants:
Hydrogenics:
HIRC:
DONG Energy:
23 November 2006, Orvel, Belgium
Michael Wenske
Lars Yde
Niels Tophøj
Hydrogenics manufactures, maintains and sells pressurised electrolysers. The plants
are made of modules with a capacity of 15Nm3/h and may be supplied in sizes up to
60Nm3/h and for a pressure of 10 bar on the hydrogen side and 8 bar on the oxygen
side.
A few years ago, Hydrogenics bought Stuart, who in 2003 had bought
Vanenborre. Stuart was founded in 1948 and with the purchase of Vandenborre the
product range was extended to comprising pressurised electrolysers. Therefore, the
technology used is the Vandenborre technology, the so-called IMET concept
(Inorganic Membrane Electrolyses Technology, marketed in 1985), which is now
marketed by Hydrogenics. The technology was originally developed by Michael
Wenske and his father based on know-how from chloride production by means of
alkali electrolysis.
Hydrogenics are developing 6 bar PEM electrolysers. However, these are not
commercially available yet.
The cell voltage of the alkali plants was specified as:
1.75V at 360mA/cm2
1.85V at 440mA/cm2
The power consumption was specified at 4.2-4.3kWh/Nm3 at 440mA/ cm2. This is an
impressingly high current density compared to IHT and Norsk Hydro, who have a
current density of 200mA/cm2 at 1.9V and 1.8V, respectively.
The tot al power consumption, including loss in transformer and rectifier as well as
deoxo-drier consumption, is 4.5-4.8kWh/Nm3. If the hydrogen must be compressed,
an additional 0.5-0.6kWh/Nm3 is consumed. With deoxo-drier, the hydrogen purity is
99.9998%.
The plants are supplied as container plants; the so-called plug and play
solutions for fully automatic operation with the possibility of remote control and
monitoring. The philosophy is that the principles of the plant must be as simple and
reliable to use as hydrogen in a bottle.
When the plant is set to standby, the pressure is maintained; but the
temperature is not. The plant will warm up again, when the production is resumed.
The dynamic range when the plant is in operation goes from 20-100%. The hydrogen
purity is 99.9% at 100% load. It is not the electrolyser that sets the limit of the control
speed; it is the power electronics. At 20% load, the oxygen content of the hydrogen is
approx 1% corresponding to 25% LEL (Lower Explosive Level). If the LEL exceeds
25%, the plant is stopped. The choice to already stop the plant at 25% LEL is due to
154
the fact that the oxygen content in case of an imbalance of the pressure between the
oxygen and the hydrogen side may develop very quickly. It is thus necessary to stop
the plant, due to the response time of the safety system.
The service life of the plants is more than 20 years at continuous operation. No
long-term experiences have been made with dynamic operation. Temperature
variations are the main cause of reduced service life, therefore it is important that the
plant is not cooled when set to standby.
The operating temperature is 75°C. This low temperature is due to the frames
in which the diaphragms are installed and which at the same time functions as seals
for the bipolar plates being made of a plastic material.
According to Michael Wenske, an operating temperature of 120°C is not
possible due to corrosion.
The dew point is -60°C for hydrogen as well as for oxygen (5ppm) and the
requirement for feed water is 5μS/cm, which is obtained by reverse osmosis and ion
exchange.
The oxygen electrodes as well as the hydrogen electrodes are made of
activated nickel grid.
The plants may be installed in non-classified rooms. You only need to
establish a ventilation outlet to fresh air in the container housing the plant. The
container is classified as zone 2.
The plants cannot be installed outdoors, because the feed water might ice up.
This means that the container must be heated, if the plant is to be set up outdoors.
The plant has separate DCS and safety system. The DCS is PLC-based, and in
case of an emergency, the safety system is able to shut down the plant to
depressurised state.
The price of electrolysers has been reduced by 50% over the past ten years. This is
due to the development of the container solution where the plants can be assembled at
the factory and thus only need to be connected to water and electricity as well as to
the customer’s hydrogen and possibly oxygen supply at the site.
The price range for Hydrogenics’ plants is around DKK 27,000/kW, which is
relatively high, but the high price range is due to the fact that Hydrogenics’ plants are
minor plants of max 250kW.
Michael Wenske estimated that for atmospheric MW plants, the price would
be reduced to DKK 4,500/kW.
Michael Wenske estimated that the total global electrolyser market or rather
the global electrolyser production capacity would be around 4000MW a year. This is
surprising, considering that IHT has an annual production capacity of 40MW.
Hydrogenics is experiencing an increasing demand for electrolysers on the
industrial market and in the energy sector. The company has approx 200 employees.
Michael Wenske is leaving Hydrogenics to start his own consultancy and project
development company within electrolysers for industrial and energy purposes.
He is going to cooperate with different manufacturers of electrolysers,
including Hydrogenics and a German company, ELT, who wishes to manufacture
non-asbestos LURGI plants. IHT, Switzerland, has the production capacity and ELT
has the know-how, but no production facilities. Their production is going to be based
solely on sub-suppliers. They have developed a non-asbestos diaphragm, but they lack
investment capital in the range of 1 million Euro. Today, the company only handles
repair and maintenance of LURGI plants.
155
6 Technical foresight
It is a common opinion that water electrolysis is an available and mature technology,
which is not wide spread because it is too expensive compared to steam reforming of
hydro carbons. As a consequence, further development has not been as high priority
as for instance fuel cells. Electrolyzers have indeed been available on truly
commercial terms and the market penetration had indeed been rather insignificant.
However, most technologies do continue developing, not despite commercial success,
but because commercial success. Cars and electronics are good examples of this. In
the introduction to this report the necessity of water electrolysis in the energy system
it is argued for, and if this holds, there will be a strong competition among
manufacturers and developers of the technology in a quite near future. Consequently,
further development of the electrolyzer technology will be crucial.
The awareness if this is growing and more research groups can be expected to
join the race for the best technology. In Denmark the authorities have been somewhat
reluctant to initiate research on water electrolysis in order not to spread the limited
resources over too much. Fuel cell research and development has already been
supported extensively for many years. However, bearing that in mind, electrolysis
development is not very different from fuel cell development and the research groups
already involved in fuel cell research have the ideal starting point for getting involved
in electrolysis (as a simplification, by just changing the direction of the current). The
electrochemistry is overall the same, although some materials problems like corrosion
will be different.
All together research and development in water electrolysis will most likely
increase in countries with tradition for electrochemical activities, especially fuel cells.
Denmark with its full-grown fuel cell R&D community will also be part of that, and
the process has already started on different levels. Risø, DTU has for some time
addressed SOEC with impressive results and other groups at the universities are
involved in catalyst development for electrolysis (Dept. of Physics, DTU), electrodes
for alkaline electrolyzers (HIH, Aarhus Univ.) and new materials for PEMEC (Dept.
Chem., DTU) to mention a few activities. Even demonstrations involving among
other things electrolyzers can now be seen (Lolland).
6.1 Alkaline electrolyzer cells (AEC)
As stated in the elsewhere in the report AEC can be regarded the standard electrolyzer
technology, but even tough it has been around for a century there will be plenty of
room for improvement. Modern technologies for especially catalyst and electrode
development have a strong potential for improving performance and lifetime. The
AEC still has the advantage that it can be produced of inexpensive materials (noble
metal free) and that oxygen kinetic is quite favourable in the alkaline environment.
The Fuel cell counterpart to the AEC has never experienced the same success
apart from niche application. This is due to the sensitivity to CO2 from the air
(carbonate formation). Carbonate formation is not a problem in the electrolyzer, as air
is not used for running the cell (oxygen is produced, not consumed).
Apart from a gradual improvement of the present cells obvious directions for
development are toward higher pressure and higher temperature. High pressure
operation is already possible with in some commercial units (32 bar), but is should be
156
possible to go further. Pressurization might well be eased by introduction of an
alkaline membrane for replacement of the porous and pressure sensitive separator.
Elevated Temperature Alkaline electrolyser Cell (ETAEC)
If the operation temperature of alkaline water electrolysis is increased from the
normal 80 - 90 °C to above 200 °C both the performance and the electricity to
hydrogen efficiency may be significantly increased.
A possible obstacle for operating at elevated temperature is the lower stability
of the materials at the increased temperature. Development of suitable materials for
the cell is thus necessary in order to develop a large scale water electrolysis plant for
operation at elevated temperatures. At present possibly suitable cell and separator
materials, which are not more expensive than low temperature alkaline electrolyzer
materials, have been identified, but the necessary long term (several years) stability
remains to be proven. Therefore, if the potential benefit of the ETAEC should be
pursued, then our recommendation is that materials test projects including accelerated
testing should be initiated.
6.2 Polymer electrolyzer cells (PEMEC)
Today the PEMEC is also more or less commercial as smaller units than the AEC.
Cost is high and need to be lowered. The traditional electrolyzer manufacturers have
become interested in the PEMEC. For instance, Norsk Hydro has now a parallel
programme for PEMEC, and others are expanding their activities in the same
direction.
The membrane makes the PEMEC well suited for pressurization, and like
mentioned above for the AEC the benefits of higher pressure and higher temperature
are also valid for PEMEC. Pressurization is to a large extent a question of engineering
of the cells, but significantly higher temperatures require other materials especially for
membrane materials.
The high temperature PEMFC (HT-PEMFC) has had a tremendous success
being developed just over the last decade. Several research groups and companies in
Denmark are now involved in HT-PEMFC research and development on all levels
from materials science to stacks and systems. It is an obvious idea to try to extent the
activities to cover also electrolyzers based on the same membrane systems. The
immediate benefits anticipated are that the excess heat produced can more easily be
utilized for heating purposes or even for steam production for steam electrolysis. As
shown in paragraph 5.2, when heat can be utilized the demand for a high electrical
efficiency is not so strong.
An interesting feature about the elevated temperature is that synthesis of
synthetic fuel might be possible in cell. Recently, methane was synthesized from CO
and H2 in a HT-PEM cell [268].
6.3 Solid oxide electrolyzer (SOEC)
During the last two years, very high initial performance of has been measured on
SOECs produced at Risø (see paragraph 4.4.3). To the author’s best knowledge, it is
the best performance ever measured on an electrolysis cell. In general, the main
reason for the high performance of the SOEC technology compared with other
electrolysis technologies is the high temperature. Both the H2O and CO2 electrolysis
157
reactions become increasingly endothermic with temperature. This makes it possible
to utilize the Joule heat produced inside the cell. The Joule heat occurs due to the
ohmic losses present in all electrolysis cells. Besides, the ohmic losses in SOECs
decreases with increasing temperature, meaning even better performance at high
temperature. More specific, the reason for the high SOEC performance of the Risø
cells is that Risø in about two decades has been among the leaders in developing the
close related SOFC technology.
This high performance makes it possible to establish a very efficient H2 and
CO production, and electrolysis on a H2O + CO2 mixture will produce synthesis gas
(H2 + CO), which can be catalyzed into various types of synthetic fuels. In such a
synthetic fuel production, some reduction in the production price may be achieved by
utilizing the heat from the catalysis reaction for steam generation.
The cost estimations show a demand for lifetimes above 3-4 years. Today, the
SOFC technology has proven lifetimes of more than 2 years, but the Risø cell does
not yet meet this lifetime demand when operated in electrolyser mode. Reasonable
stability over more than 1000 h has been achieved at current densities of 0.5 A/cm2 or
below. Ongoing research and development addresses this issue. The glass sealings
used to avoid the electrode gasses to mix with surrounding gasses is partly made of
SiO2. Large amounts of Si has been detected in the Ni/YSZ electrode on tested
SOECs by EDX mapping in a SEM. Test with pre-treated glass sealings show
significant improvements.
The recommendation is that the following main subjects should be addressed in the
future R&D: 1) precise identification of the mechanism of the cell degradation, 2)
developments of highly durable cells, 3) further feasibility studies through cell and
stack testing, 4) construction of pressurized cell and stack test facilities, 5)
construction of prototype electrolyzer systems, 6) more detailed technical and
economical modelling should be done parallel to the experimental work.
6.4 Reversible fuel cells
The concept of using the same cells as both fuel cells and electrolyzer cells is very
tempting. Many attempts have been made over the years to develop so called
“reversible” or bi-functional cells, but it appears to be a challenge to match the good
fuel cells and good electrolyzers with the same cell. The best starting point is
probably the electrolyzer, and not the fuel cell, because the electrolyzer faces the
highest voltages and thus the most challenging conditions with respect to corrosion.
Standard PEMFC electrodes based on carbon materials are not durable in electrolyzer
mode (at least not on the oxygen side), but there is no reason why electrolyzer
electrode cannot operate in fuel cell mode. However, as a special case, the solid oxide
electrolyzer cells at Risø are developed from the equivalent fuel cells, but in this case
the materials are based on oxides, i.e. materials that are already oxidized in the first
place. In conclusion, development of reversible fuel cells is to a large extent
electrolyzer technology.
158
6.5 System development
In practical application the electrolyzer stack is not standing alone. A system around
needs development as well, and when water electrolysis is to be used for energy
conversion this system will be different to a system for production of industrial gasses
and the possible interplay with other parts of the energy system now becomes
important. Questions to ask can be
• How can the produced oxygen be utilized?
• How can the excess heat (if any) be utilized?
• Will the produced hydrogen be used for further fuel synthesis?
• Do we need to store hydrogen (and oxygen) for later back conversion?
Oxygen can be use din an oxy-fuel combustion process resulting in high temperatures
and highly concentrated CO2 in the flue gas. This CO2 can be useful for the synthesis
of synthetic fuels, as there will be no need for N2 removal.
The heat produced must be at a high enough temperature for sufficient heat
transport through heat exchangers. Heat transport always requires temperature
differences as the driving force. Transport of heat from a 60-80ºC electrolyzer to a
district heating net is not possible without additional electrical energy spent in a heat
pump. If the temperature is 150ºC or more it might be practical.
Synthesis of methanol, methane or other synthetic fuels might be desirable
because it eases storage and later fuelling. It should be studied under which condition
the electrolyzer itself or the system can facilitate this synthesis. Generally, atomic
hydrogen and oxygen (or their respective ions) that appear at the electrolyzer
electrodes are more reactive than the molecular hydrogen and oxygen they form
before they are released as products of the electrolysis process.
In case hydrogen is produced with the aim of storing, say wind energy, the
oxygen produced should be stored as well. The production ratio (1:2) is of course the
same as needed for the back conversion, and fuel cells operated on pore oxygen
(instead of air with only 21% oxygen) performs with a higher efficiency.
Finally, the Danish wind industry has for a long time been in the lead world wide.
This position might be difficult to maintain in the future as large countries with cheap
labour (India, China) are now taking up the competition. A decisive parameter in the
global competition could be supporting technologies optionally to go with the wind
turbines. When wind power is enhanced to cover a larger fraction of the energy
supply in other countries that Denmark and a few more, technology to handle the
produced hydrogen will be asked for.
159
7 Adaptation of electrolysis in the Danish energy
system
7.1 Electrolysis in the Danish power system
The occurrence in recent times of serious blackouts in America and Europe
underscores the fact that additional measures are urgently needed to avoid such costly
incidents.
In addition integration of increasing generation capacity from renewable
energy sources is a challenge to the operation of the system.
More effective generation side management and demand side management
services will play an important role in meeting current and future need for reliability.
The need and potential for integrating energy storage or energy conversion in
electrical power systems with high wind penetration is already widely recognised
within electric power utilities [269].
In this context electrolysers - being both flexible electric loads, energy
conversion systems and storage - can increase the flexibility of the system and be an
important measure to allow the integration of additional renewable energy in the
Danish power system.
7.2 The power supply of electrolysers
As mentioned in section 3.1.2, the power supply of large-scale electrolyser systems
will be from a three-phase, high-voltage line. To convert this into the low-voltage dc
power needed for the electrolyser cell, a combination of transformer and rectifier unit
is usually used.
7.2.1 The load control
The individual electrolyser cells require voltage adjusted in the interval 80-100% as to
change the current from 20% to 100% (Figure 7-1). The Maximum DC-voltage is 920
V [270].
Figure 7-1. Voltage variation as function of load current variations.
160
7.2.2 Intermittent operation
An interesting system aspect of electrolytic processes for the production of hydrogen
is the possibility of load management in electric grids. Like all electrochemical energy
converters electrolysers can respond to load changes almost instantaneously. Highly
dynamic electrolysers can thus be used for hydrogen production in the case of
fluctuating production from wind power plants. In combination with a hydrogen
storage tank, an electrolyser can be used for load management in the same way as a
variable electricity consumer. This load management can lead to a higher utilization
of wind power plants, since electrolysers can be used to balance the grid.
7.2.3 AC/DC conversion and power regulating
Electrolysers operates at direct current (DC) at about 1,6 to1,9V volt per cell. The grid
is operated at 400 volt or more, alternating current (AC) at 50 Hertz (Hz). Therefore
the current (electricity) from the grid have to be rectified before it can be used to
power the electrolyser.
The voltage level also has to be adapted to the level required by the
electrolyser by a transformer.
If it is required that the plant can be regulated continuously from zero to full
power, a unit for this purpose is also necessary. The drawing below shows the
diagram fore such a power supply.
Figure 7-2. Power supply for an alkaline electrolyser, utilizing a standard inverter for
induction motors, to regulate the power consumption by the plant.
161
Figure 7-3. Hydrogen from water electrolyses used for production of synthetic fuels.
7.2.4 Converters and converter configurations
Thyristor technology is normally used for electrolysis. Alternatively, rectifier can be a
diode type. Such industrial power rectifiers are a well known and employed in various
applications within chemical and metallurgical industries.
Rectifier transformer
Rectifier transformers are designed for combination with a diode or thyristor
rectifiers. The transformers may have a built-in or separate voltage regulation unit for
direct output regulation of diode rectifiers, and correspondingly a power factor
improvement with a thyristor rectifier.
The thyristor converter requires a transformer with only coarse (stepped)
voltage regulation, if any. This will often be done by a no-load tap changer (NLTC).
If a diode rectifier is applied, this will require a continuous regulation of the
secondary voltage of the transformer. This is done by a combination of a stepped
voltage regulation, applying a coarse- or multicoarse-fine on-load tap changer
(OLTC), and the fine tuning voltage regulation by applying saturable reactors on the
LV side of the transformer [271].
Rectifier
The basic 3-phase rectifier is the so-called 6-pulse converter. In order to improve
harmonic emissions on plant electrical supply it is normal to combine two such 6pulse converters in parallel, to produce a 12-pulse converter. One converter is phase
shifted by 30 degrees with respect to the other one. This can be continued to 24 pulses
or higher. Most industrial rectifiers use 12 pulse converters, with a few larger systems
using 24-pulse converters [272]. Two 12 pulse converters can be configured as a 24
pulse system, with one 15 degrees phase shifted from the other, or they can be a
configured as a 12 pulse system, with both converters in phase.
The obvious reason for employing thyristor technology in high power
applications with low DC-voltage and very high DC-currents is that standard thyristor
rectifiers offers the highest efficiencies compared e.g. modern IGBT technology.
162
Disadvantages of thyristor rectifiers are:
(a)
Reactive power consumption, as the power factor on the AC supply side will
vary with firing angle of the thyristor bridge.
(b)
Disturbances on the AC supply side in terms of frequency harmonics and
commutation voltage sags
High power electrolyser system will thus need additional equipment for power factor
correction by means of e.g. capacitor banks and tuned filters for suppression of
critical harmonics.
Auxiliary systems
High power rectifiers auxiliary equipment includes cooling system. Depending on
rating, location and ambient conditions rectifier cooling can by air or water.
7.2.5 Alternative converter configurations
The above-described concepts employ so-called load-commutated rectifiers.
Alternative converter technology using self-commutated semiconductors may
improve the performance, not at least on the grid side.
Two alternatives of AC/DC converters, PWM converters and DC choppers,
are described in the following.
PWM converter applications
PWM (pulse width modulated) converters using GTO thyristors or IGBT transistors
have generally not been applied for high power converter systems. The technology
has been widely used for motor drives and similar applications and has also been
introduced for HVDC transmissions systems.
One advantage of PWM converters is that by applying space vector control of
the converter it is possible to control the active power (or the DC voltage) and the
reactive power on the AC side independently. Two quadrant operations allowing two
ways of reactive power flows may offer possibilities in terms of ancillary services to
the power system as described in the following.
Another advantage is the reduced low order harmonics generated on the AC
supply side. High frequency harmonics with orders around multiples of the switching
frequencies will, however, be generated and may need additional filtering.
The most obvious drawback of PWM converters is the losses. In addition to
the conduction losses PWM converters have significant switching losses, which are
directly proportional to switching frequency.
However, a remarkable development has been seen in the development of
semiconductors in terms of power and voltage capabilities. This has been seen in the
development of variable speed motor drives for industry and power applications, the
employment of power electronics in wind turbines and also development of high
voltage converters for distribution and transmission application.
163
Based on this it could possibly be expected that the performance and capabilities of
these converters would make it attractive also for high current applications like
electrolysers.
In addition the value of the control capabilities may eventually become higher
and thus make PWM type of converters more cost effective.
DC chopper applications
Systems employing a diode rectifier on the supply side and a chopper in the DC
circuit are currently used e.g. in the metallic industries for e.g. dc arc furnace
applications but can also be used in heavy current applications in the electrochemical
industry [273], and likewise for electrolysis.
Main benefits of chopper applications are fast dynamic response; high line
power factor over the entire power range without power factor correction; minimal
harmonic distortion of the ac power feeder without the use of harmonic filters; high
system efficiencies over the total output power range; reduction in overall system size
and cost.
Fast dynamic response is hardly required in electrolysis systems; however, the
remaining benefits may be worthwhile considering.
7.3 Grid connection and integration in the electrical power system
Integration of electrolysis can be done in various ways. Technically depending on the
size of plants and the voltage level it connects to.
Integration in the electricity market basically of course is to operate as a
flexible load by cost effective production of Hydrogen). However the system and the
market also have a demand for various ancillary services, to which electrolysis may
offer its capabilities.
7.3.1 Large scale electrolysers at transmission levels
Large-scale electrolyser as e.g. envisaged by DONG Energy’s in the range of several
hundreds of MW’s will have to connect at the transmission level at 132-150 kV.
Connection of large-scale electrolysers in the transmission system can mitigate
transfer capacity problems in the transmission system, which occurs e.g. in periods
with high wind generation. In this way significant costs of transmission line upgrades
can be avoided.
In addition electrolysers may offer possibilities for improving the security of
the system by e.g. including them in remedial action schemes as significant load that
can be disconnected.
Finally, as will be discussed in the following, by applying PWM converter
technology electrolysers may eventually add significantly to the reactive power and
voltage control of the system.
164
7.3.2 Medium and small scale electrolysers at distribution levels
Medium scale electrolysers (50 MW or less) may connect to the distribution system at
voltage levels 50-60 kV or 10-20 kV.
Here it is of interest to consider Energinet.dk’s development plans for a new
network structure in Denmark. It’s a concept based on a two-layer structure where the
150 kV and the 400 kV transmission levels are to be jointly planned and operated and
the local grids below each 150/60 kV transformer station will constitute network cells
in which monitoring and control are performed. The cell structure is illustrated in
Figure 7-4.
165
Figure 7-4. Cell structure in a future grid design (Energinet.dk)
Basically, the new strategy requires a new set of functional requirements to the cells
for controlling the reactive power locally, controlling disconnection of production or
consumption in critical situations as to enable immediate compensation of suddenly
emerged imbalances.
The long-term target of Energinet.dk’s cell structure is to prepare the power
system operation with a major part of the generation in the distribution networks, not
least the increased renewable energy resources.
A local control structure will monitor and control the power flow. When
considering the active power, a cell should be self-regulating, but not necessarily selfsufficient. Emergency conditions should be handled within the cell offering the
possibility of island operation and black start.
Installation of electrolysers in distribution network fits well with these plans
and allows increased control of active and reactive power in the network cells. This
will not the least be the case if plans for retrofitting existing on-shore wind turbines
and installing new larger wind turbines. In this case the total distributed wind capacity
will increase and the need for local reserves will increase if curtailments should be
avoided.
The electrolysers may provide significant value both as a flexible load
responding to market signals, but also as a provider of ancillary services (active and
reactive power control).
7.3.3 Electrolysers combined with wind generation
Several research projects have already on a lab scale investigated how electrolysers
can be operated in small systems together with wind generation.
For a large-scale implementation of electrolysers it is likely that they would be
combined with wind power in a portfolio operated on the market by a common
balance responsible party. In this context it is unimportant whether the plants have
different points of common coupling (PCC) with the power system as indicated in
Figure 7-5 below or they are directly interconnected to a common PCC.
166
Obviously if the units are of significant ratings and located far apart, local problems
of transfer capacity can arise.
Portfolio
PCC
Wind farm
Power System
Electrolyser
Power pool
Balance responsible party
Figure 7-5. Combination of wind generation and electrolyser in a common portfolio.
If the electrolyser should participate in reactive power or voltage control and this
should be combined with the wind generation, they should be connected to same
PCC.
The benefit of this combination first of all is that the intermittent renewable
production can be reduced. Converting peak production to hydrogen and storing it for
off-peak use can increase the capacity factor and value of the renewable energy.
Similarly the use of the transmission systems can be improved since it need
not be designed for transfer of peak productions, which may occur in few hours.
Ultimately, hydrogen production in combinations with wind can also help
utilize wind energy resources in areas where new transmission lines cannot be built.
7.4 Flexible electrical load
Flexible electrical load FEL is defined as variations of consumer load on a short-term
basis as response to price signals.
The price signal form the market can be
•
•
•
The price from spot market
Real time prices from reserve and balance markets
Individual prices or signals agreed between consumer and power producer
Apart from regular consumers, FEL also includes conversion of electricity to other
energy sources.
Large consumer can already now purchase electricity on the spot market and
can thus adapt their consumption to the prices.
In order to make the market more flexible, new forms of contracts may
develop, which will make the price flexibility simpler.
167
The Danish TSO estimates that the potential for price flexibility of regular
consumers in Denmark is approximately 660 MW [274]. The integration of
electrolysis plants of say 100-200 MW would thus add significantly to the flexibility.
7.4.1 The electricity price variations
The electricity price depends on the supply and demand in the system. As wind power
enters the market at low marginal costs, the wind generation is seen to have a
significant impact on the price as indicated in Figure 7-6 below.
800
700
Price [DKK/MWh]
600
500
400
300
200
100
0
0
500
1000
1500
2000
2500
Wind generation [MW]
Figure 7-6. Electricity prices in DK-West in 2006 dependant of wind generation.
Data from Energinet.dk [275].
Looking at the daily variation of electricity prices, wind can expectedly be seen to
have a major impact in low peak hours as indicated in Figure 7-7 below.
Figure 7-7. Average daily electricity price variation in 2006 at different levels of
wind generation. Data from Energinet.dk [275].
Increasing wind capacity in the system will tend to give larger fluctuations in
electricity prices and increase problems of power overflow when wind power can
168
cover the complete load. A more flexible demand will mitigate these problems and
give more stable prices.
Electricity
price
Supply
Demand
Flexible
demand
More stable
prices
Large
fluktuation
in prices
Wind
MWh
Figure 7-8. Wind power influence on electricity prices with and
without flexibility in demand.
7.4.2 Electrolysis as flexible load
In principle there are no technical restrictions for electrolysers to operate as a flexible
load and purchase power from the spot market
The value of integrating electrolysers into the system is demonstrated in the
following.
More or less as consequence of electricity prices depending on wind
generation, net exports tend to increase with increased wind generation. This is
illustrated by the duration curve of wind generation and exports for the Western
Denmark put together as shown in Figure 7-9. The net exports obviously depend on
other factors and the curve is an average filtered for these variations. In any case the
graph indicates that a significant part of the wind generation in DK-West is exported
and on the whole that the international market works.
169
Wind generation and net exports from DK-West 2006
2.500
DK-Vest: Wind
Net export (avaraged)
2.000
MW
1.500
1.000
500
0
0
1000
2000
3000
4000
5000
6000
7000
8000
Hours
Figure 7-9. Wind generation and net exports DK-West 2006.
Data from Energinet.dk [275].
However, from a local Danish power system perspective this is not necessarily an
advantageous situation. Not the least because wind generation is subsidised. An
increased integration of electrolysers (and other flexible loads, storage and energy
conversion systems) would increase the value of the wind generation.
7.5 Ancillary services
Various definitions of ancillary services exist [276- 278]. Presently mainly generating
units supply ancillary services. However, the utilisation of dump loads by means of
thermal storage has been introduced.
2
Electrolysis systems may depending on type of electrolyzer and type of converter
configuration offering ancillary services to the electrical power system in terms of:
•
•
•
•
•
Primary active power/frequency control
Automatic emergency power/frequency control
Secondary active power control (automatic or manual reserves)
Dynamic reactive power and voltage control
Reactive power reserves
The possibilities of electrolysers to ad value to the system are discussed in the
following. This may be relevant for stand-alone electrolysers operating in the system
and market as well as electrolysers operating in a portfolio including renewable
generation and/or conventional generation.
170
7.6 Active power control
7.6.1 Active power control on electrolysers
The choice of electrolyser type has influence on the power control capabilities of the
system. Dynamic operation requires an advanced control of the cooling system. The
cooling must be controlled according to the power consumed by the electrolyser and
not only by the temperature at the outlet cooling water.
Standard alkaline electrolysers presently in operation are not designed for fast
control. Large electrolysers like Norsk Hydro are designed for continuous operation
and take approx. two hours to start-up and increase production to 100% [270].
Increased cooling can decrease the start-up time. Small alkaline electrolyser
cells are available, which offer 20 seconds start-up time.
Downward regulation of the load can be achieved within few seconds.
A Solid Oxide Electrolyser system can be controlled by regulation of the temperature
(by changing the current through the cells) and at the same time control the water or
steam supply. The thermal control of the cells will be important for the control of
power to the electrolyser. Downwards regulation from 100% to 0% can happen in
approx. 30 seconds, whereas the upward control can take 15 minutes or more.
PEM electrolysers will offer better performance in terms of control range and
regulation time.
A dynamic range of the PEM electrolyser 5–100 % of rated capacity has been
achieved and typically the electrolyser will have a response time from 5-100 % in less
than a second.
In this context the performance of electrolysers in combination with wind
generation is of interest. As indicated in the table below, this combination offers a
variation of possibilities that can ensure fast response
System need
Increase power in the
system
Electrolyser
Wind
Decrease load
Fast response
< 5 seconds
Available when in
operation
Increase load
Slow response
Available when in
operation
Increase generation
Only possible to in
special cases with
prior reduction
Decrease generation
Fast response
< 5 seconds
Available when in
operation
Figure 7-10. Power control possibilities for a combination of wind and electrolyser
Decrease power in the
system
171
7.6.2 Primary active power/frequency control
The demand for primary power/frequency control in the electrical power system is
defined in UCTE and NORDEL requirements.
The primary power/frequency control needs to be provided automatically
within 5-30 seconds.
If the electrolyzer would enter this marked segment, it would require that the
electrolyser would have to be in operation all times and continuously offer margins
for the required control range.
Not all present electrolysers on the marked can offer the performance required
without change in the control and cooling system.
7.6.3 Automatic emergency reserves
It is presently required that generation units (both conventional power plants and wind
farms), which are connected to the transmission system, are obliged to provide
automatic active power reserves for the support of the power system in case of
disturbances. The control should be activated automatically when frequency
excursions outside specified limits occur. This is not a service that will be paid for
with reference to the Electricity Act.
Presently such obligations are not applied to specific flexible loads. However,
in case emergency reserves are not sufficient to stabilise the power system after the
disturbance, frequency protections will automatically shed loads (i.e. disconnect
consumers).
Electrolysers may offer such services and as indicated in Figure 7-10 excellent
performance can be achieved in combination with wind generation.
Automatic control activated by a measured frequency deviation can easily be
applied.
In this context it is also interesting to look at the Danish TSO long-term plans
to restructure the power system and dividing it into more or less autonomous network
cells as described in section 7.3.2. An increase in installed wind power capacity in
Denmark on-shore, which is envisaged, would expectedly develop as distributed
generation rather than large wind farms, and would thus connect into the distribution
systems. In this case the combination with electrolysers will increase the possibilities
of achieving the emergency reserves needed in the cell
7.6.4 Secondary active power control (automatic or manual reserves)
Secondary reserves are in demand daily to balance deviations, which occur due to
prediction errors (load and renewable generation) and breakdowns.
This is an area where the market has extended and incitement to consumers to
participate is introduced. For a consumer two services can be offered
(a) Upward regulation by disconnection or reduction of load
(b) Downward regulation by increasing load.
172
The activation of the regulation can be automatic based on a remote control set point
from the network regulator or done manually.
Manual reserves (regulating power) can be offered in two ways
(a) An availability obligation under which the consumer is obliged provide of a
specified amount of regulating power for specified periods and receive availability
compensation as well as energy payment.
(b) Placing regulating power bids when attractive, freely select offered quantity and
price, and receive energy payment when the bid is activated.
Presently bids have to be minimum 10 MW and maximum 50-75 MW. Manually
reserves have to be activated within 10-15 minutes.
The market for regulating power is a segment where an electrolyser will be
able to participate. A large electrolyser may offer a percentage of the capacity as
regulating power and operate the remaining capacity as a flexible load on the spot
market.
As described above in 7.6.1 electrolysers may meet the requirements for
activation time and the ramping time. Here start-up and ramp-up time of alkaline
electrolysers are the only critical point. Technical solutions time may need to be
developed.
How an electrolyser may work on the market depends on how the system and
the market develop in the future with a significant increase in wind generation.
An availability obligation may be entered for downward regulation (increase
load) on condition that the start-up time and power ramping up can meet the
activation time requirement. An obligation for upward regulation (decrease load) may
be less attractive, since this would require continuous operation of the offered
capacity throughout high price peak hours.
Presently the agreed capacities and the availability compensation for
downward regulation (increase load) tends to be significantly lower that for up-ward
regulation.
New wind farms are expected to have improved performance in terms of
power control and will need to accept some balance responsibility.
Both stand-alone electrolysers and electrolysers in combination with wind
generation will have good possibilities to operate on the free market for regulating
power. Considering the fact that electrolysers are almost certain to be in operation
during periods with high wind, which are also the periods where the system has
highest needs for regulating power, makes it favourable both for the owners and the
power system.
7.7 Reactive power and voltage control
7.7.1 Reactive power control of electrolysers
Basically electrolysers are consumers and as described in section 7.2 the power
supply be means of diode or thyristor rectifiers will implicit result in a consumption of
reactive power which has to be compensated for by installing capacitor banks.
To improve performance, alternative converter configurations may be
considered as described in section 7.2.5.
173
The most obvious solution would be to employ diode rectifier and a chopper
for power control. This application would keep reactive power consumption at low
level (constant high power factor) but will not offer any reactive power control
capabilities.
Ultimately, PWM converters would as mentioned offer the possibility of
controlling of active power (or the DC voltage) and reactive power on the AC side
independently.
Electrolysers only need one way of active power flow and a converter
designed for two quadrant operations may be applied. The reactive power capability
may look like Figure 7-11 below.
Figure 7-11. PQ-capability curve for HVDC Light technology modified to show twoquadrant operation only (Source: ABB: “HVDC, A “firewall” against disturbances in
high-voltage grids”.
This shows that the electrolyser may provide wide possibilities for reactive power
control to the power system, even at low load.
As mentioned, it may not be obvious to apply PWM converters to high current
applications like electrolysers due to losses. But in the long run the development in
performance and capabilities of PWM converters or the development (increase) in
power system need of voltage control may result in a situation where break-even is
met.
174
7.7.2 Reactive power reserves, voltage control and short circuit level
In nowadays power system conventional generating units are the main source of
reactive power and voltage control. All units in operation are obliged to participate,
however a minimum number of power stations need to be in operation to provide the
voltage control and short circuit level that guarantee stable and secure operation of the
power system.
If local reactive power reserves are insufficient in certain areas of the grid, the
system operator mitigates problems by installing additional active compensation, e.g.
by means of rotating synchronous condensers or static var-compensators (SVC’s).
Presently there is no market for other suppliers of reactive power reserves.
Wind farms for instance are required to have to reactive power regulation, but are
generally simply asked to operate at neutral power factor.
The response of voltage control, which is needed by the power system, can be divided
in
(a) Voltage control under normal operation (small-signal response)
(b) Voltage control during grid faults (large-signal response)
Different response times are required in the two cases.
An electrolyser supplied from a PWM converter and employing state of the art
controls will offer a performance similar to an SVC and - taking into account the PQcapability chart - will be able to offer both kinds of voltage control.
An SVC has been installed in a 132 kV substation Radsted on Lolland to
compensate for voltage fluctuations from Nysted off shore wind farm. This has a
regulating range +85/-65 Mvar. The total costs of this including substation etc. is
approx. 100 million DKK.
This might be an incentive to develop electrolysers in this direction.
175
8 Conclusion
If the challenges pinpointed by the IPCC and others shall be met tremendous changes
in our energy system is mandatory over a limited period of time. Strongly increased
use of renewables requires large scale energy conversion and most likely techniques
for large scale storage. There are two reasons for this. 1) The renewable energy
production is typically very fluctuating (e.g. wind, solar) and the production will not
be able to match the demand if a large fraction of the energy supply is fluctuating like
the wind power. 2) The transport sector will for a long time need a fuel which can be
stores onboard vehicles and ships. Battery powered vehicles are under steady
development, by they still have a long way to go especially in terms of range before
they are flexible enough for replacing fuelled vehicles. For heavy transport like trucks
the demands are more severe, and for ships and planes battery systems are even more
speculative. Even with a future introduction of a certain fleet of battery vehicles, the
demand for some sort of fuel will be inevitable.
Fuel can to some extend be produced from biomass and waste, but these
sources will only partly cover the need. Some countries plan on depending on nuclear
power, and this way it might be easier to match the overall energy demand, but it
doesn’t solve the problem of the transport sector as the energy still has the form of
electricity or heat, just like the renewables (apart from biomass and waste).
In conclusion, conversion from electricity to a fuel is inevitable in a future
energy system, and no matter if this fuel is pure hydrogen or synthetic fuels
(methanol, methane, gasoline etc.) the first step is production of hydrogen from water
splitting. Although there are a number of methods for water splitting, at present the
only realistic technique for this process at a large scale is water electrolysis.
8.1 Electrolyzer technology
There are basically 3 types of electrolyzers of practical interest, namely the alkaline
electrolyzer (AEC), the polymer electrolyzer (PEMEC) and the solid oxide
electrolyser (SOEC).
The AEC is in its present for the standard technology which is has been
available commercially for many years up to the MW size. It is manufactured from
rather inexpensive materials without noble metals and the working temperature is
below 100ºC. The AEC technology appears to be quite bulky and gives the
impression of having been developed for low cost manufacturing rather than
sophistication. However, further development of the AEC technology is envisioned.
With operating temperatures up to about 200ºC improved performance is expected
and the possibility of utilizing the excess heat increases. Another route for
improvement is to increase the working pressure, which is also the delivery pressure
for the hydrogen and oxygen. If the product are to be used for further synthesis at high
pressure or stored in pressure tanks this pre-compression reduced the subsequent
compression work. The ideal work of compression from 1 to 10 bar is equivalent to
the ideal work of compression from 10 to 100 bar so; even a moderate pressurization
of the electrolyzer is advantageous. Today large scale 32 bar AEC’s are available.
The PEMEC is in its initial commercial state, but the units are so far smaller
than the AEC’s. They comprise a more compact and sophisticated technology based
176
on noble metal catalysts. Several of the companies know for their AEC’s are these
years establishing parallel development lines for PEMEC. The fact that the electrolyte
is practically solid (in contrast to the AEC) is an advantage if the system is to be
pressurized. The working temperature for PEMEC’s is below 100ºC like for the
conventional AEC, but it is also here suggested to increase the temperature to the
interval 100-200ºC just like it has recently been done for the PEM fuel cell. This will
require a high temperature membrane system. The benefits expected are similar to
those for AEC at elevated temperature.
A common challenge for increasing the working temperature for the two
systems is that corrosion effects will be more pronounced and consequently,
significant materials research is necessary.
The SOEC is working at high temperatures in the range of 700-1000ºC. It benefits
from the fact that the electrical energy theoretical required decreases with temperature
(a tendency that is also beneficial when the temperature is increase for the low
temperature systems mentioned above, although the effect is much more pronounced
at SOEC temperatures). The SOEC is not yet close to commercialization, but its
potential for operations at the thermo-neutral voltage makes it a very promising future
candidate with high electric efficiency.
It is found that H2 and CO can be produced at attractive production costs,
using SOECs. The H2 production cost was found to be 71 US¢/kg equivalent to 30
$/barrel crude oil using the HHV. The CO production cost was found to be 5.6
US¢/kg equivalent to 34 $/barrel crude oil using the HHV. If heat for steam
generation can be provided from a waste heat source, the production price can be
lowered even further.
For lifetimes above 3-4 years the H2 production price starts to become
insensitive to the life time. For the CO production price this is about 6 years. The
production cost was found to be lowest at ETn. Here the efficiency from electricity to
fuel was found to be 93% for CO production and 96% for H2 production. These
figures do not include heat loss to the surroundings.
Recycling of CO2 for carbon dioxide neutral synthetic fuels can be performed
with calcium carbonate, which is a well-known CO2 absorbent and has been
suggested for CO2 sequestration. Thermodynamic equilibrium calculations show on
the other hand that CO2 capture/recycling using magnesium carbonate can be
operated at approximately 400 ºC lower than the 800 ºC for calcium carbonate.
A carbonate cycle for CO2 capture/recycling is definitively technically
feasible, but the practical and economic aspects regarding calcium carbonate,
magnesium carbonate or calcium magnesium carbonate have to be assessed to
determine the most suitable absorbent for CO2 capture/recycling.
8.2 Large commercial electrolyzers
A review of the performance of commercially available electrolysers was carried out.
It covered solely AEC systems because the PEMECs on the market were too small for
large scale hydrogen production. Five western and one eastern company related to
delivery of electrolyser systems have been identified. Prices of the western plants
show that systems not smaller than 1 MW have to be used in order to get reasonable
specific prices (EURO/kW).
177
The companies were:
• Norsk Hydro, Norway(ambient pressure: 200 to 2000 kW, pressurized 50 to 300
kW
• Hydrogenics, Belgium/Canada (pressurized:60 to 240 kW)
• Iht, Switzerland (ambient pressure: 14 to 1500 kW, pressurized: 500 to 3400 kW)
• AccaGen, Switzerland (pressurized: 7 to 500 kW)
• Erre Due, Italy (pressurized: 100 to 200 kW)
• Uralhimmash, Russia. (ambient pressure: 1520 to 3150 kW, pressurized: 201250kW)
Electrolysers have the reputation of being very expensive. It is true but often when the
price pr. kW of a specific electrolyser is mentioned the size of the plant is not given.
The specific price of electrolysers (EURO / kW) is strongly dependent of the size of
the plant. The price analysis below shows it very clearly. The prices of electrolyzers
for which prices were obtained are plotted in Figure 8-1. It can be seen that the price
per kW installed capacity vary with a factor of 10 dependent of the size of the plant.
Euro / kW
Prices of Alcaline Electrolysers
10000
9000
8000
7000
6000
5000
4000
3000
2000
1000
0
0
500
1.000
1.500
2.000
2.500
3.000
kW
1 Bar, HYDRO
4 Bar, H2Industrial
16 Bar, HYDRO
6 Bar, AccaGen
10 Bar, Hydrogenics
1 Bar ELT
Figure 8-1. Prices of electrolyzers as a function of production capacity
The prices are collected over the last 5 years from year 2000 and therefore not
consistent. Anyhow, the graphic shows clearly that in order to obtain relatively cheap
elecrtolysers, they have to be as large as possible and not smaller then about 1 MW.
The efficiency of an electrolyser is defined as the ratio of the higher heat value
(HHV) of the hydrogen produced and the DC electricity consumption of the
electrolyser. The commercial electrolyzers have an electricity consumption of 4.1 to
4.8 kWh per normal m3 produced. Using the HHV of 3.5 kWh/m3 hydrogen, the
efficiencies can be calculated to between 85 and 73%.
Hydrogen is produced from water and electricity. The water consumption is
about one litre per normal cubic metre (Nm3) of hydrogen and the electricity required
is approx 4kWh per Nm3. This means that the water price is minimal compared to the
price of the electricity.
178
The electricity cost for hydrogen production was calculated on the following
basis: 1) Actual electricity prices through 2006 (varying). 2) The price of the largest
electrolyzer in the study. Assumptions: 1) Electrolyzer depreciation over 10 years,
75% efficiency (HHV). The production cost for hydrogen was then varying between
0.45 and 0.50 kr/kWh for between 25 and 100% use of the electrolyzer (part time
production at lowest electricity prices).
If, moreover, the value of oxygen and heat was included in the calculation the
cost was between 0.35 and 0.40 kr/kWh.
If the electrolyzer efficiency is assumed td to be 100% instead of 75% the cost
is between 0.30 and 0.33 kr/kWh in the same utilisation interval. This limited
reduction is not an argument against research and development of more efficient
electrolyzers, but a very strong indication that there is absolutely no reason to await
more efficient electrolysers to start business development.
8.3 Future research and development
Research and development in water electrolysis will most likely increase in countries
with tradition for electrochemical activities, especially fuel cells. Denmark with its
full-grown fuel cell R&D community will also be part of that, and the process has
already started on different levels. Risø, DTU has for some time addressed SOEC
with impressive results and other groups at the universities are involved in catalyst
development for electrolysis (Dept. of Physics, DTU), electrodes for alkaline
electrolyzers (HIH, Aarhus Univ.) and new materials for PEMEC (Dept. Chem.,
DTU) to mention a few activities.
The development of the three different electrolysers (AEC, PEMEC and
SOEC) will be a steady gradual improvement of key figure in performance and
lifetime. New promising routes are mainly higher working temperatures of the two
first systems, and higher pressures for all systems.
A higher working temperature results in lower theoretical demand of energy in
the form of electricity (work) and more in the form of heat. As heat is always
produced to some extent due to internal losses, the result is improved electrical
efficiency. Another advantage of higher temperature is that excess heat developed can
better be utilized, e.g. in the district heating system provided that the temperature is
high enough. If the excess heat is utilized it is not a waste and the overall efficiency is
high even though the electrical efficiency is not.
Materials for AEC at elevated temperature are identified, but testing especially
with respect to stability is needed. For the development of PEMEC at higher
temperatures it is obvious to start with the high temperature membranes already
applied for fuel cells and modify the electrodes to avoid corrosion.
Higher working pressure of the cell makes it possible to deliver hydrogen at
higher pressure and consequently, less compression work is needed for subsequent
compression for storage in pressure tanks. Commercial electrolyzer working at 32 bar
are already available, but is should be possible to go further.
The production of synthetic fuels inside or in direct connection with the electrolyzer is
possible at elevated temperature. In SOEC not only water electrolysis is possible CO2
can also be electrolyzed to CO which in combination with hydrogen for synthesis gas,
the starting point for Fischer-Tropsch synthesis of organic fuels. Methane synthesis
has recently been carried out in PEM cells at elevated temperature.
179
The concept of using the same cells as both fuel cells and electrolyzer cells is very
tempting. Many attempts have been made over the years to develop so called
“reversible” or bi-functional cells, but it appears to be a challenge to match the good
fuel cells and good electrolyzers with the same cell. The best starting point is
probably the electrolyzer, and not the fuel cell, because the electrolyzer faces the
highest voltages and thus the most challenging conditions with respect to corrosion.
Standard PEMFC electrodes based on carbon materials are not durable in electrolyzer
mode (at least not on the oxygen side), but there is no reason why electrolyzer
electrode cannot operate in fuel cell mode. However, as a special case, the solid oxide
electrolyzer cells at Risø are developed from the equivalent fuel cells, but in this case
the materials are based on oxides, i.e. materials that are already oxidized in the first
place. In conclusion, development of reversible fuel cells is to a large extent
electrolyzer technology.
In practical application the electrolyzer stack is not standing alone. A system around
needs development as well, and when water electrolysis is to be used for energy
conversion this system will be different to a system for production of industrial gasses
and the possible interplay with other parts of the energy system now becomes
important.
Oxygen can be used in an oxy-fuel combustion process resulting in high temperatures
and highly concentrated CO2 in the flue gas. This CO2 can be useful for the
synthesis of synthetic fuels, as there will be no need for N2 removal.
The heat produced must be at a high enough temperature for sufficient heat
transport through heat exchangers. Heat transport always requires temperature
differences as the driving force. Transport of heat from a 60-80ºC electrolyzer to a
district heating net is not possible without additional electrical energy spent in a heat
pump. If the temperature is 150ºC or more it might be practical.
Synthesis of methanol, methane or other synthetic fuels might be desirable
because it eases storage and later fuelling. It should be studied under which condition
the electrolyzer itself or the system can facilitate this synthesis. Generally, atomic
hydrogen and oxygen (or their respective ions) that appear at the electrolyzer
electrodes are more reactive than the molecular hydrogen and oxygen they form
before they are released as products of the electrolysis process.
In case hydrogen is produced with the aim of storing, say wind energy, the
oxygen produced should be stored as well. The production ratio (1:2) is of course the
same as needed for the back conversion, and fuel cells operated on pore oxygen
(instead of air with only 21% oxygen) performs with a higher efficiency.
The leading position of the Danish wind industry might be difficult to maintain in the
future as large countries with cheap labour (India, China) are now taking up the
competition. A decisive parameter in the global competition could be supporting
technologies optionally to go with the wind turbines. When wind power is enhanced
to cover a larger fraction of the energy supply in other countries that Denmark and a
few more, technology to handle the produced hydrogen will be asked for.
Finally, to conclude in short form: with a large fraction of renewables in the energy
system water electrolysis is inevitable even though the technology is not perfect.
Efficiencies, short term cost and the advantages and drawbacks of the different
180
technologies can be, and will be, discussed. However, a more fundamental question
could be “what is the alternative, if business as usual is not an option?”
181
9 References
[1]
Momirlan M, Veziroglu TN. Current status of hydrogen energy. Renew Sustain Energy
Rev., 2002;6:141–79.
[2]
Rothstein J. Hydrogen and fossil fuels. Int. J. Hydrogen Energy 1995;20:283–6.
[3]
Goswami DY, Mirabal ST, Goel N, Ingley HA. A review of hydrogen production
technologies. In: Shah RK, Kandilkar SG, editors. Fuel cell science, engineering and
technology, Proceedings of the first international conference on fuel cell science,
engineering and technology. Rochester, NY: ASME; April 21–23, 2003. p. 61–74.
[4]
Bockris JO’M. Energy options. New York: Halsted Press; 1980.
[5]
Kreuter W, Hofmann H. Electrolysis: the important energy transformer in a world of
sustainable energy. Int J Hydrogen Energy 1998;23:661–6.
[6]
Shinnar R, Shapira D, Zakal S. Thermochemical and hybrid cycles for hydrogen
production. A differential comparison with electrolysis. Ind Eng Chem Process Des
Dev 1981;20:581 –93.
[7]
Alexander Volta, Philosophical Transactions of the Royal Society of London, Vol. 90.
(1800), pp. 403-431.
[8]
P.-O. Nissen. Poul la Cour og vindmøllerne (in Danish). Polyteknisk Forlag, Lyngby,
Denmark , 2003, ISBN 87-5020951-5.
[9]
Electrochemical Hydrogen Technologies, H. Vendt, Elsevier, Amsterdam (1990).
[10] S. C. Singhal, in: Proceedings of the Sixth International Symposium, ed. S. C. Singhal
and M. Dokiya (The Electrochemical Society, Pennington, 2007) 39-51.
[11] A. V. Virkar, K. Fung, and S. C. Singhal, in: Ionic and Mixed Conducting Ceramics,
ed. The Electrochemical Society, Pennington, 1998) 113-124.
[12] Wendt H, Plzak H. Hydrogen production by water electrolysis. Kerntechnik 1991; 56
(1) pp. 22–8.
[13] R. L. Leroy and A. K. Stuart, Advanced unipolar electrolysis, Int. J. Hydrogen Energy,
Vol. 6, No. 6, pp. 589-599, 1981.
[14] Divisek J. Water electrolysis in low- and medium-temperature regime. In: Wendt H,
editor. Electrochemical hydrogen technologies-electrochemical production and
combustion of hydrogen. Oxford: Elsevier, 1990.
[15] M. BRAUN, Description of processes for producing hydrogen by advanced
electrolysis, status of development, Symposium on Hydrogen in Air Transportation,
Stuttgart (September 1979).
[16] J. EVANGELISTA, I. PHILLIPS and L. GORDON, Electrolytic Hydrogen Production:
An Analysis and Review, NASA Technical Memorandum TM X-71856 (1975).
[17] A. P. FICKETr and F. R. KALHAMMER, Water electrolysis, in Hydrogen: Its
Technology and Implications, Vol. 1 (Ed. by K. E. Cox and K. D. Williamson), CRC
Press Inc., Cleveland (1977).
[18] A. GAN, Hydrogen production by water electrolysis (translation of European Space
Agency Report No. DLR-Mitt 7439), NTIS Report ESA TF-250 (1976).
182
[19] A. MENTH and S. STUCKI, Present state and outlook of the electrolytic H2production route, in Hydrogen Energy System, Vol. 1, (Ed, by T. N. Veziroglu and W.
Seifritz), p. 55, Pergamon Press, Oxford (1978).
[20] R. L. LERov and A. K. STUART, Unipolar water electrolysers: A competitive
technology, in Hydrogen Energy System, Vol. 1, (Ed. by T. N. Veziroglu and W.
Seifritz), p. 359, Pergamon Press, Oxford (1978).
[21] R. L. LERov and A. K. STUART, Present and future costs of hydrogen production by
unipolar water electrolysis, in Industrial Water Electrolysis (Ed. by S. Srinivasan, F. J.
Salzano and A. R. Landgrebe), p. 117, Proceedings Vol. 78-4, The Electrochemical
Society Inc., Princeton (1978).
[22] A. K. STUART, Modern electrolyser technology in industry, American Chemical
Society Annual National Meeting, Symposium on Non-Fossil Chemical Fuels, Boston
(1972).
[23] A. K. STUART, Electrolyser technology for hydrogen oxygen production, Symposium
Course on Hydrogen Energy Fundamentals, Miami Beach (1975).
[24] Ullmanns Enzyklopadie der technischen Chemie (Ullmanns Encyclopaedia of
Technical Chemistry), Vol.18 (1967).
[25] H. Wendt, H. Hofmann and V. Plzak, Anode and cathode-activation, diaphragmconstruction and electrolyzer configuratiuon in advanced alkaline water electrolysis,
Int. J. Hydrogen Energy. Vol. 9, No. 4, pp. 297-302, 1984.
[26] J. Fischer, H. Hofmann, G. Luft and H. Wendt, AIChE J. 26, 794 (1980).
[27] G. Kissel, F. Kucesa, C. R. Davidson and S. Srinivasan, Proc. Symp. Industrial Water
Electrolysis 78, 222 (1978).
[28] J. W. VOGT, Materials for electrochemical cell separators, Report NAS CR72493,
NASA Lewis Research Center, December (1968).
[29] W. B. CeNDALL & Y. HARADA, Fuel cell matrix studies, Report AD761 512.
Prepared by I.I.T. Research Institute for U.S. Air Force Aerospace Research
Laboratories, May (1973).
[30] P. J. MORAN & G. E. STONER, A proposed failure mechanism of asbestos separator
material, Proceedings of the ERDA Contractor's Review Meeting on Chemical Energy
Storage and Hydrogen Energy Systems, Airlie, Virginia, November (1976).
[31] P. GODIN, R. GRAZIOtti, A. DAMPEN & P. MASNIERE, Int. J. Hydrogen Energy 2,
291 (1977).
[32] J. M. GRAS & J. J. LeCoz, Asbestos corrosion study in hot caustic potash solution.
Silicate ions influence on electrode overvoltages, Proceedings of the 2nd World
Hydrogen Energy Conference, Zurich, Switzerland, p. 255, August (1978).
[33] R. W. VINE & S. T. NARSAVAGE, Porous matrix structures for alkaline electrolyte
fuel cells, Abstract No.45, Electrochemical Society Extended Abstract, 75, 96 (1975).
[34] Y. HARADA, A. Z. HED & W. B. CRANDALL, Chemical protection of asbestos.
U.S. Patent No. 3,891,461 (1975).
[35] (a) M. KLEIN & R. ASTRIN, Hydrogen-oxygen regenerative fuel cells, NASA Lewis
Research Center, Report NASA CR-1244 (1969).
(b) M. KLEIN & R. ASTRIN, Hydrogen-oxygen regenerative fuel cells, NASA Lewis
Research Center, Report NASA CR-1683 (1970).
183
[36] A. P. MEYER, R. W. VINE & S. T. NARSAVAGE, Development of advanced fuel
cell system (phase IV), Appendix A, Porous matrix structures for alkaline electrolyte
fuel cells, United Technologies Corporation, Report NASA CR 135 030, January
(1976).
[37] R. E. POST, Evaluation of potassium titanate as a component of alkaline fuel cell
matrices, NASA Lewis Research Center, Report NASA TN D-8341, November (1976).
[38] (a) Laboratoire de Marcoussis, Compagnie Generale d'Electricite, Nouveaux
catalyseurs et separateurs pour l'e1ectrolyse de l'eau: Premiere optimisation des
parametres de fonctionnement, from Seminar on hydrogen as an energy vector: Its
production, use and transportation, ECC Report EUR6085, Brussels, p. 166, October
(1978).
(b) A. J. APPLEBY & G. CREPY, Improvements in electrolysis technology in alkaline
solution, Proceedings of the 2nd World Hydrogen Energy Conference, Zurich,
Switzerland, Vol. 1, pp. 227-240, August (1978).
(c) R. L. VIC & J. P. POMPON, Amelioration de procede e1ectrolytique de fabrication
d'hydrogene, from Seminar on Hydrogen as an Energy Vector, ECC Seminar Preprints,
Brussels, p. 163, February (1980).
[39] L. H. BAETSLE, 8f. D. HUYS, Inorg. J. Nucl. Chem. 30, 639 (1968).
[40] H. VANDENBORRE & R. LEYSEN, Electrochimica Acta 23,803 (1978).
[41] H. VANDENBORRE, R. LEYSEN, J. P. TOLLENBOOM & L. H. BAETSLE,
Alkaline inorganic-membraneelectrolyte water electrolysis, in Seminar on Hydrogen as
an Energy Vector: Its Production, Use and Transportation, ECC Report EUR6085,
Brussels, p. 161, October (1978).
[42] H. VANDENBORRE, L. H. BAETSLE, W. HEBEL, R. LEYSEN, H. NACKAERTS
& G. SPAEPEN, Development and parametric testing of alkaline water electrolysis
cells for hydrogen production based on inorganicmembrane- electrolyte technology, in
Hydrogen as an Energy Vector, EEC Seminar Preprints, Brussels. p.143,
February(1980).
[43] H. VANDENBORRE, R. LEYSEN, L. H. BAETSLE, Int. J. Hydrogen Energy 5, 165
(1980).
[44] E. HAUSMANN, Amelioration of asbestos-based diaphragms and the possibility of
their replacement with semi-permeable membranes, in Seminar on Hydrogen as an
Energy Vector: Its Production, Use and Transportation, EEC Report EUR 6085,
Brussels, p. 229. October (1979).
[45] P. PERROUD, Rev. gen. Electr. 85,627 (1976).
[46] P. PERROUD & G. TERRIER, The use of porous metallic diaphragms for hydrogen
mass production with alkaline water electrolysis, Proceedings of the 2nd World
Hydrogen Energy Conference, Zurich, Switzerland, Volume 1, pp. 241-245, August
(1978).
[47] J. FISCHER, H. HOFFMAN, C. LUFT & H. WENDT, Fundamentals and
technological aspects of medium temperature (MT) high pressure (HP) water
electrolysis, in Hydrogen as an Energy Vector: Its Production, Use and Transportation,
ECC Report EU6085, Brussels, p. 277, October (1978).
[48] J. DIVISEK, J. MERGEL & H. SCHMITZ, Improvement of water electrolysis in
alkaline media at intermediate temperatures, Proceedings of the 3rd World Hydrogen
Energy Conference, Tokyo, Japan, Vol. 1, pp. 209-221, June (1980).
184
[49] Batelle Geneva Research Centre, Technical Note. Plasma-sprayed inorganic
membranes as diaphragms for electrolysis (1978).
[50] J. N. MURRAY, Long term stability of thermoplastics and fiber reinforced
thermoplastics in alkaline solution, oxygen environments, Abstract No. 21,
Electrochemical Society Extended Abstracts, Vol. 76-1, p. 54 (1976).
[51] J. N. MURRAY & M. R. YAFFE, Aqueous caustic electrolysers at high temperatures,
Abstract No. 483, Electrochemical Society Extended Abstracts, Vol. 78-1, p. 1209
(1978).
[52] J. N. MURRAY, M. R. YAFFE, Improvements leading to high efficiency alkaline
electrolysis cells, Proceedings of the 3rd World Hydrogen Energy Conference, Tokyo,
Japan, Vol. 1, p. 15, June (1980).
[53] D. R. DREGER, The polysulfones: heat resistant, superstrong, and ultrastable, Machine
Design, January 12 (1978).
[54] (a) Union Carbide, Guide to chemical resistance of Udel polysulfone, Technical Data
Sheets F-46706, August (1977).
(b) Union Carbide, Radel TM polyphenylsulfone, the tough high temperature
thermoplastic, Technical Data Sheet F-46264, October (1976).
(c) I.C.I., Polyethersulfone, the high temperature engineering thermoplastic, Technical
Data Sheet P.E.S. 10.
[55] Minnesota Mining and Manufacturing Co., Matrix construction for fuel cells, British
Patent No. 1,435,420, April 4 (1974).
[56] J. BOURGANEL, Procede de preparation de membranes anisotropes semi-permeables
en polyarylethersulfones, Brevet Canadien No. 1,009,419, May (1977).
[57] Phillips Chemical Company, Ryton polyphenylene sulfide resins: physical, chemical,
thermal and electrical properties, TSM-266, March (1977).
[58] J.-C. SOHM & L. MAS, New diaphragms for electrolysers, from Seminar on Hydrogen
as an Energy Vector, ECC Seminar Preprints, Brussels, p. 261, February (1980).
[59] (a) Japan's Sunshine Project, Summary of hydrogen energy R & D: Studies on
production of hydrogen, Agency of Industrial Science and Technology, January (1979).
(b) E. TORIKAI, Y. KAWAMI, M. NAMBA & S. SHUHEI, Chemical-resistant
diaphragm, Get. Often. 2,717,512, November 10 (1977).
(c) E. TORIKAI, K. YOUJI, K. SHINICHIRO, N. MUTSUSUKE & S. SHUHEI,
Method of making a chemicalresistant diaphragm thereof, U.S. Patent No. 4,111,866,
September 5 (1978).
[60] N. WAKABAYASHI, E. TORIKAI, Y. KAWAMI & H. TAKENAKE, Advanced
alkaline water electrolysis, Proceedings of the 3rd World Hydrogen Energy
Conference, Tokyo, Japan, Vol. 1, p. 59, June (1980).
[61] F. J. SALZANO, Hydrogen production and storage in utility systems, Semi-Annual
Progress Report, July, 1975 to December 31, 1975, BNL 50590, January (1976).
[62] R. S. YEO, J. MCBREEN, G. KISSEL, F. KULESA & S. SRINIVASAN,
Perfluorosulfonic acid (Nation) membrane as separator for advanced alkaline water
electrolyser, submitted for publication in J. appl. Electrochem. BNL Report 27136.
[63] P. M. SPAZIANTE & L. GIUFFRE, Studio della struttura ottimale di un elettrolita
solido polimerico per l'elettrolisi dell' acqua, presented at Seminar on Hydrogen as an
Energy Vector, Brussels, February (1980).
185
[64] L. GIUFTRE & C. MODICA, Diaframmi e membrane per l'elettrolisi dell'acqua, from
"Seminar on Hydrogen as an Energy Vector", ECC Seminar Preprints, Brussels, p. 243,
February (1980).
[65] H. VOGEL & C. S. MARVEL, J. Polymer Sci. 1,511 (1961).
[66] E. C. CHENEVEY, Ultrafine polybenzimidazole fibers, Celanese Research Co., NASA
CR 159644, June (1979).
[67] P. M. HERGENROTHER & H. H. LEVINE, J. Polymer Sci. 5,453 (1967).
[68] J. V. DUFFY & W. P. KILROY, Silver oxide-zinc battery separator development,
Abstract No. 54, The Electrochemical Society Extended Abstracts, 76, 146 (1976).
[69] J. E. FRENCH & R. J. SCHWARZ, An organic resin for high temperature applications,
Abstract No. 38, The Electrochemical Society Extended Abstracts 75, 82 (1975).
[70] M. H. Miles, Electroanal. Chem. Interfacial Electrochem.,60, 89 (1975).
[71] C. Ballleux, A. Damien, and A. Montet, Int. J. Hydrogen Energy, 8, 529 (1983).
[72] P. W. T. Lu and S. Srinivasan, J. Appl. Electrochem., 9, 269 (I979).
[73] R. L. LeRoy, M. B. I. Janjua, R. Renaud, and U. Leuenberger, J. Electrochem. Soc.,
126, 1674 (1979).
[74] P. W. T. Lu and S. Sriuivasan J. Electrochem. Soc., 125, 1416 (1978).
[75] M. H. Miles, G. Kissel, P. W. T. Lu, and S. Srinivasan, J. Electrochem. Soc., 123, 332
(1976).
[76] A. C. C. Tseung and S. Jasem, Electrochim. Acta, 22, 31 (1977).
[77] H. C. Angus, D. E. Hall, and R. D. Giles, in "Proceedings of the 4th World Hydrogen
Energy Conference," Pasadena, CA (1982).
[78] G. Fiori, C. M. Mari, B. Perra, L. Vago, and P. Vitali, in "Proc. Hydrogen as an Energy
Vector, Comm. Eur. Communities," p. 223, (1980).
[79] D. P. Gregory, K. F. Blurton, and N. P. Biederman, in "Industrial Water Electrolysis,"
S. Srinivasan, F. J. Salzano, and A. R. Landgrebe, Editors, p. 54, The Electrochemical
Society Softbound Proceedings Series, Princeton, N.J. (1978).
[80] A. J. Appleby, G. Crepy, and J. Jacquelin, Int. J. Hydrogen Energy, 3, 21 (1978).
[81] J. M. Gras and' M. Pernot, in "Electrode Materials and Processes for Energy
Conversion and Storage," J. D. E. McIntyre, S. Srihivasan, and F. G. Will, Editors, p.
425, The Electrochemical Society Softbound Proceedings Series, Princeton, N.J.
(1977).
[82] M. H. Miles, Electroanal. Chem. InterJacial Electrochem., 60, 89 (1975).
[83] D. E. Hall, Electrodes for Alkaline Water Electrolysis, J. Electrochem. Soc., Vol. 128,
No.4, 740 (1981).
[84] D. E. Hall, J. Electrochem. Soc., 129, 310 (1982).
[85] N. Wakabayashi, E. Torikai, Y. Kawami, and H.Takenaka, Adv. Hydrogen Energy 2,
Hydrogen Energy Progress I, 59 (1981).
[86] J. Balajka, Int. J. Hydrogen Energy, 8, 755 (1983).
[87] G. E. Stoner and P. J. Moran, in "Proceedings of the Symposium on Industrial Water
Electrolysis," S. Srinivasan. F. J. Salzano, and A. R. Landgrebe, Editors, p. 169, The
Electrochemical Society Softbound Proceedings Series, Princeton. NJ (1978).
186
[88] G. L. Cahen, Jr., P. J. Moran, L. L. Scribner, and G. E. Stoner, This Journal, 128, 1877
(1981).
[89] G. E. Stoner and P. J. Moran, in "Industrial Water Electrolysis," S. Srinivasan, F. J.
Salzano, and A. R. Landgrebe, Editors, p. 169, The Electrochemical Society Softbound
Proceedings Series, Princeton, NJ (1978).
[90] A. C. C. Tseung, S. Jasem, and M. N. Mahmood, in "Industrial Water Electrolysis," S.
Srinivasan, F. J. Salzano, and A. R. Landgrebe, Editors, p. 461, The Electrochemical
Society Softbound Proceedings Series, Princeton, NJ (1978).
[91] J. Divisek and H. Schmitz, Int. J. Hydrogen Energy, 7, 703 (1982).
[92] P. Ragunathan, S. K. Mitra, and M. G. Nayar, ibid., 6, 487 (1981).
[93] M. Bernard. L. Mas, M. Prigent, and J. C. Sohm, in "Proc. Hydrogen as an Energy
Vector, Comm. Eur. Communities," p. 283 (1980).
[94] M. Prigent and T. Nenner, Adv.Hydrogen Energy 3 ,Hydrogen Energy Progress IV,
299 (1982).
[95] H. Wendt and V. Plzak, Electrochim. Acta, 28, 27 (1983).
[96] L. Martin, J. Diette, M. Prigent, M. Bernard, J. Demarsy, and C. Sellier, Comm. Eur.
Communities Final Report 502-78-1-EHF (1981).
[97] R. D. Giles, Adv. Hydrogen Energy 3,Hydrogen Energy Progress IV, 279 (1982).
[98] A. Garat and J. Gras, Int. J. Hydrogen Energy, 8, 681 (1983).
[99] A.J. Salkind and U. Falk, "Alkaline Storage Batteries," John Wiley, New York (1969).
[100] T. Shirogami, K. Murata, and M. Ueno, Abstract 27, The Electrochemical Society
Extended Abstracts, Vol. 81-2, Denver, Colorado, Oct. 11-16, 1981.
[101] D. E. Hall, J. Electrochem. Soc., 130, 317 (1983).
[102] A. C. C. Tseung, S. Jasem, and M. N. Mahmood, in "Proceedings of the Symposium on
Industrial Water Electrolysis,"
[103] S. Srinivasan, F. J. Salzano, and A. R. Landgrebe, Editors, p. 161, The Electrochemical
Society Softbound Proceedings Series, Pennington, NJ (1978).
[104] G. Bronsel and J. Reby, gZectrochim. Acta, 25, 973 (1980).
[105] P. W. T. Lu and S. Srinivasan, Journal of the Electrochemical Society, 125, 1416
(1978).
[106] P. Rasiyah and A. C. C. Tseung, Adv. Hydrogen Energy 3, Hydrogen Energy Progress
IV, 383 (1982).
[107] M. C. M. Man, S. Jasem, K. L. K. Yeung, and A. C. C. Tseung, "Sere. Hydrogen as an
Energy Vector, Comm. Eur. Communities," p. 255 (1978).
[108] P. Rasiyah, A. C. C. Tseung, and D. B. Hibbert Journal of the Electrochemical Society,
129, 1724 (1982).
[109] P. Rasiyah and A. C. C. Tseung, Journal of the Electrochemical Society, 130, 2384
(1983).
[110] H. Vandenborre, R. Leysen, H. Nackaerts, and Ph. VanAsbroeck, Adv. Hydrogen
Energy 3, Hydrogen Energy Progress IV, 107 (1982).
[111] P. Rasiyah and A. C. C. Tseung, Journal of the Electrochemical Society, 130, 365
(1983).
187
[112] G. Fiori, C. Mandelli, C. M. Mar/, and P. V. Scolari, Adv.Hydrogen Energy 1,
Hydrogen Energy Systems I, 193 (1979).
[113] H. Vandenborre, R. Leysen, and H. Nackaerts, Int. J.Hydrogen Energy, 8, 81 (1983).
[114] M. Prigent, L. J. Mas, and F. C. Verillon, in "Proceedings of the Symposium on
Industrial Water Electrolysis," S. Srinivasan, F. J. Salzano, and A. R. Landgrebe,
Editors, p. 234, The Electrochemical Society Softbound Proceedings Series,
Pennington, NJ (1978).
[115] A. C. C. Tseung, Journal of the Electrochemical Society, 125, 1660 (1978).
[116] H. Obayashi and T. Kudo, Mater. Res. Bull., 13, 1409 (1978).
[117] Y. Matsumoto, S. Yamada, T. Hishida, and E. Sato, Journal of the Electrochemical
Society, 127, 2360 (1980).
[118] T. Otagawa and J. O'M. Bockris, Journal of the Electrochemical Society, 129, 2391
(1982).
[119] J. O'M. Bockris, T. Otagawa, and V. Young, J. Phys. Chem., 150, 633 (1983).
[120] J. O'M. Bockris and T. Otagawa, J. Phys. Chem., 87, 2960 (1983).
[121] J. O'M. Bockris and T. Otagawa, Journal of the Electrochemical Society, 131, 290
(1984).
[122] P. F. Carcia, R. D. Shannon, P. E. Bierstedt, and R. B. Flippen, Journal of the
Electrochemical Society, 127, 1974 (1980).
[123] M. DeKay Thompson and A. L. Kaye, Trans. Electrochem. Soc., 60, 229 (1931).
[124] G. Grube and W. Gaupp, Z. Elektrochem., 45, 290 (1939).
[125] J. M. Gras and M. Pernot, in "Proceedings of the Symposium on Electrode Materials
and Processes for Energy Conversion and Storage," J. D. E. McIntyre, S. Srinivasan,
and F. G. Will, Editors, p. 424, The Electrochemical Society Softbound Proceedings
Series, Pennington, NJ (1977).
[126] D. E. Hall, U.S. Pat. 4,240,887, 1980.
[127] A. Nidola, P. M. Spaziante, and L. Giuffre, in "Proceedings of the Symposium on
Industrial Water Electrolysis," S. Srinivasan, F. J. Salzano, and A. R. Landgrebe,
Editors, p. 102, The Electrochemical Society Softbound Proceedings Series,
Pennington, NJ (1978).
[128] P. W. T. Lu and S. Srinivasan, Abstract 328, The Electrochemical Society Extended
Abstracts, Vol. 76-1, Washington, DC, May 2-7, 1976.
[129] "Hydrogen Storage and Production in Utility Systems," F. J. Salzano, Editor,
Brookhaven Nat'l Lab. Report 50472. 1975.
[130] T. Osaka, Y. Iwase, H. Kitayama, and T. Tchino, Bull. Chem. Soc. Jpn. 56, 2106
(1983).
[131] M. H. Miles, Y. H. Huang, and S. Srinivasan, Journal of the Electrochemical Society,
125, 1931 (1978).
[132] H. S. Horowitz, J. M. Longo, and H. H. Horowitz, Journal of the Electrochemical
Society, 130, 1851 (1983).
[133] W.A. Fischer and D. Janke, Die Anwendbarkeit von ZrO2-Y2O3-ZrO2-CaO und ThO2 Y2O3 Festelektrolyten bei 1000o-1600 oC). (The Practicability of ZrO2-Y2O3-ZrO2- CaO
and ThO2- Y2O3 Solid Electrolytes at 1000°-1600 °C), WV, Opladen, 1972.
188
[134] J.E. Mrochek, "The Economics of Hydrogen and Oxygen Production by Water Electrolysis and Competitive Processes," Abundant Nuclear Energy , Washington 1969,
S.107-122.
[135] L.J. Nuttall, A.P. Fickett and W.A. Titterington, "Hydrogen Generation by Solid
Polymer Electrolyte Water Electrolysis," GEC, Direct Energy Conversion Programs,
Lynn, Massachusetts 01910.
[136] L.H. Russell, L.J. Nut tall and A.P. Fickett, "Hydrogen Generation by Solid-Polymer
Electrolyte Water Electrolysis," paper presented to the American Chemical Society,
Division of Fuel Chemistry, Chicago, August 1973.
[137] Oberlin R, Fischer M. Status of the membrane process for water electrolysis. In:
Veziroglu T, Getoff N, Weinzierl P, editors. Hydrogen energy progress VI, proceedings
of the sixth world hydrogen energy conference. Oxford: Pergamon Press; 1986. p. 333–
40.
[138] Pahomov VP, Fateev VN. Electrolysis of water with solid polymer electrolyte. RRC
“Kurchatov Institute”, 1990 (Preprint, in Russian).
[139] Millet P, Andolfatto F, Durand R (1996) Design and performance of a solid polymer
electrolyte water electrolyzer. Int J Hydrogen Energy 21:87–93
[140] Friedland RJ, Speranza AJ (1999) Integrated renewable hydrogen utility system.
Proceedings of the 1999 US DOE Hydrogen Program Review, NREL/CP-570-26938
[141] Prince-Richard S, Whale M, Djilali N (2005) A techno-economic analysis of
decentralized electrolytic hydrogen production for fuel cell vehicles. Int J Hydrogen
Energy 30:1159–1179
[142] www.hydroelectrolysers.com
[143] http://www.hydrogenics.com/
[144] http://www.protonenergy.com
[145] http://www.ginerinc.com
[146] Davenport RJ, Schubert FH. Space water electrolysis: space station through advanced
missions. J Power Sources, 1991;36:235–50.
[147] B. V. Tilak, P. W. T. Lu, J. E. Colman, and S. Srinivasan, in "Comprehensive Treatise
of Electrochemistry," Vol. 2, J. O'M. Bockris, B. E. Conway, E. Yeager, and R.E.
White, Editors, pp. 1-104, Plenum Press, New York (1981).
[148] (a) L. J. Nuttall, in "Proceedings of the 1st World Hydrogen Conference (WHEC),"
Miami, Florida, D. 6B-21 (1976);
(b) L. J. Nuttall and J. H. Russell, in "Proceedings of the 2nd WHEC," Zurich,
Switzerland, Vol. 1, p. 391 (1978);
(c) J. H. Russel, in "Proceedings of the 3rd WHEC," Tokyo, Japan, Vol. 1, p. 3 (1980);
(d) General Electric Company, "Solid Polymer Electrolyte Water Electrolysis
Technology Development for Large-Scale Hydrogen Production," Final Report, Oct.
1977-Nov. 1981, DOE/ET/262020-1 (1981).
[149] (a) S. Stucki and R. Muller, in "Proceedings of the 3rd WHEC," Tokyo, Japan, Vol. 4,
p. 1799 (1980); (b) R. Oberlin and M. Fischer, in "Proceedings of the 6th WHEC,"
Vienna, Austria, Vol. 1, p. 333 (1986).
[150] (a) H. Takenaka and E. Torikai, Kokai Tokyo Koho, Jap. Patent 5,538,934 (1980);
189
(b) H. Takenaka, E. Torikai, Y. Kawami, and N. Wakabayashi, Int. J. Hydrogen
Energy, 7, 397 (1982);
(c) H. Takenaka, E. Torikai, Y. Kawami, N. Wakabayashi, and T. Sakai, Denki
Kagaku, 52, 351 ((1984);
(d) H. Takenaka, E. Torikai, Y. Kawami, and T. Sakai, ibid., 53, 261 (1985);
(e) H. Takenaka, Y. Kawami, I. Uehara, N. Wakabayashi, and M. Motone, ibid., 57,
145 (1989).
[151] Marshall A, Borresen B, Hagen G, Tunold R, Tsypkin M. Development of oxygen
evolution electrocatalysts for proton exchange membrane water electrolysis.
Proceedings of the First European Hydrogen Energy Conference, 2–5 September, 2003,
Grenoble, France. p. 45.
[152] Rasten E, Hagen G, Tunold R. Electrocatalysis in water electrolysis with solid polymer
electrolyte. Electrochim Acta 2003;48:3945–52.
[153] Tanaka Y, Uchinashi S, Saihara Y, Kikuchi K, Okaya T, Ogumi Z. Dissolution of
hydrogen and the ratio of the dissolved hydrogen content to the produced hydrogen in
electrolyzed water using SPE water electrolyzer. Electrochim Acta 2003;48:4013–9.
[154] Mitsugi C, Harurni A, Kenzo F. WE-NET: Japanese hydrogen program. Int J Hydrogen
Energy 1998;23:159–64.
[155] Yamaguchi M, Horiguchi M, Nakanori T. Development of large-scale water
electrolyzer using solid polymer electrolyte in WE-NET. Proceedings of the 13th
World Hydrogen Energy Conference, vol. 1, June 12–15, 2000, Beijing, China. p. 274–
81.
[156] Hashimoto A, Hashizaki K, Shimizu K. Development of PEM water electrolysis type
hydrogen production system for WE-NET. Proceedings of the 14th World Hydrogen
Energy Conference on CD, June 9–13, 2002, Montreal, Canada.
[157] Wan N, Pang Z. Advanced in study of solid polymer electrolyte water electrolysis.
Proceedings of the 13th World Hydrogen Energy Conference, vol. 1, June 12–15, 2000,
Beijing, China. p. 266–8.
[158] Shao Z, Yi B, Han M. The membrane electrodes assembly for SPE water electrolysis.
Proceedings of the 13th World Hydrogen Energy Conference, vol. 1, June 12–15, 2000,
Beijing, China. p. 269–73.
[159] Millet P, Durand R, Pineri M. Preparation of new solid polymer electrolyte composites
for water electrolysis. Int J Hydrogen Energy 1990;15:245–50.
[160] http://www.hsssi.com
[161] http://www.protonenergy.com
[162] http://www.h-tec.com
[163] Millet P, Alleau T, Durand R. Characterization of membrane-electrode assemblies for
solid polymer electrolyte water electrolysis. J Appl Electrochem 1993;23:322–8.
[164] Michas A, Millet P. Metal and metal oxides based membrane composites for solid
polymer water electrolysers. J Membrane Sci 1991;61:157–62.
[165] Takenaka H, Sakai T, Kawami Y, Torikai E. Effects of surface roughening of Nafion_
on electrode plating, mechanical strength, and cell performances for SPE water
electrolysis. J Electroanal Chem 1990;12:3777–84.
190
[166] Kondoli M, Yokoyama N, Inazumi C, Maezawa S, Fujiwara N, Nishimura Y, Oguro K,
Takenaka H. Development of solid polymer-electrolyte water electrolyser. J New
Mater Electrochem Sys 2000;3:61–6.
[167] Ledjeff K, Heinzel A, Peinecke V, Mahlendorf F. Development of pressure
electrolyser and fuel cell with polymer electrolyte. Int J Hydrogen Energy
1993;19:453–9.
[168] Fateev VN, Archakov OV, Lyutikova EK, Kulikova LN, Porembsky VI. Electrolysis of
water in systems with solid polymer electrolyte. Russ J Electrochem 1993;29(4):551–7.
[169] Grigoriev SA, Khaliullin MM, Kuleshov NV, Fateev VN. Electrolysis of water in a
system with a solid polymer electrolyte at elevated pressure. Russ J Electrochem
2001;37(8):953–7.
[170] Marshall A, Børresen B, Hagen G, Tsypkin M, Tunold R. Nanocrystalline IrxSn1-xO2
electrocatalysts for oxygen evolution in water electrolysis with polymer electrolyte—
effect of heat treatment. J New Mater Electrochem Sys 2004;7:197–204.
[171] Marshall A, Børresen B, Hagen G, Tsypkin M, Tunold R. Preparation and
characterisation of nanocrystalline IrxSn1-xO2 electrocatalytic powders. Mater Chem
Phys 2005;94:226–32.
[172] Marshall A, Børresen B, Hagen G, Tsypkin M, Tunold R. Electrochim. Acta
2006;51:3161–7.
[173] Marshall A, Børresen B, Hagen G, Tsypkin M, Tunold R. Nanocrystalline IrxSn1-xO2
electrocatalysts for oxygen evolution in water electrolysers using proton exchange
membranes. In: 55th ISE meeting, Thessaloniki, Greece, 2004. p. 732.
[174] Marshall A. Electrocatalysts for the oxygen evolution electrode in water electrolysers
using proton exchange membranes: synthesis and characterisation. PhD thesis.
Trondheim, Norway: NTNU; 2005.
[175] Rasten E. Electrocatalysis in water electrolysis with solid polymer electrolyte. PhD
thesis. Trondheim, Norway: NTNU; 2001.
[176] Beer H. Improvements in or relating to electrodes for electrolysis. British Patent
1,147,442.
[177] Trasatti S. Electrocatalysis in the anodic evolution of oxygen and chlorine. Electrochim
Acta 1984;29:1503–12.
[178] Kötz R, Stucki S. Stabilization of RuO2 by IrO2 for anodic oxygen evolution in acid
media. Electrochim Acta 1986;31:1311–6.
[179] Hutchings R, Mu¨ller K, Stucki S. A structural investigation of stabilized oxygen
evolution catalysts. J Mater Sci 1984;19:3987–94.
[180] Roginskaya Y, Morozova O, Loubnin E, Popov A, Ulitina Y, Zhurov V, et al. X-ray
diffraction, transmission electron microscopy and X-ray photoelectron spectroscopic
characterization of iridium dioxide+tantalum oxide (Ta2O5) films. J Chem Soc Faraday
Trans 1993;89:1707–15.
[181] Trasatti S. Physical electrochemistry of ceramic oxides. Electrochim Acta
1991;36:225–41.
[182] De Battisti A, Brina R, Gavelli G, Benedetti A, Fagherazzi G. Influence of the valve
metal oxide on the properties of ruthenium based mixed oxide electrodes: Part I.
Titanium supported RuO2/ Ta2O5 layers. J Electroanal Chem 1985; 200:93 –104.
191
[183] A. Marshall, B. Børresen, G. Hagen, M. Tsypkin, R. Tunold, Hydrogen production by
advanced proton exchange membrane (PEM) water electrolyzers—Reduced energy
consumption by improved electrocatalysis, Energy 32 (2007) 431–436.
[184] R. L. Leroy, Int. J. Hydrogen Energy, 8, 401 (1983).
[185] R. L. Leroy, C. T. Bowen and D. J. Leroy, J. Electrochem. Soc., 127, 1954 (1980).
[186] R. L. Leroy, J. Electrochem. Soc., 130, 2158 (1983).
[187] K. Onda, T. Kyakuno, K. Hattori and K. Ito, J. Power Sources, 132, 64 (2004).
[188] R. L. Leroy, M. B. I. Janjua, R. Renaud and U. Leuenberger, J. Electrochem. Soc., 126,
1674 (1979).
[189] Electrochemical hydrogen technologies: Electrochemical production and combustion of
hydrogen, H. Wendt, Editor, Elsevier, Amsterdam, The Netherlands, (1990).
[190] S. Rausch and H. Wendt, J. Electrochem. Soc., 143, 2852 (1996).
[191] www.hydro.com/electrolysers, (2007).
[192] A. C. Ferreira, E. R. Gonzalez, E. A. Ticianelli, L. A. Avaca and B. Matvienko, J.
Appl. Electrochem., 18, 894 (1988).
[193] M. H. Miles, G. Kissel, P. W. T. Lu and S. Srinivasan, J. Electrochem. Soc., 123, 332
(1976).
[194] Y. Ogata, H. Hori, M. Yasuda and F. Hine, J. Electrochem. Soc., 135, 76 (1988).
[195] T. Ohta, J. E. Funk, J. D. Porter and B. V. Tilak, Int. J. Hydrogen Energy, 10, 571
(1985).
[196] M. H. Miles, J. Electroanal. Chem, 60, 89 (1975).
[197] N. V. Krstajic, B. N. Grgur, N. S. Mladenovic, M. V. Vojnovic and M. M. Jaksic,
Electrochim. Acta, 42, 323 (1997).
[198] P. W. T. Lu and S. Srinivasan, J. Electrochem. Soc., 125, 265 (1978).
[199] P. Dabo, H. Menard and L. Brossard, Int. J. Hydrogen Energy, 22, 763 (1997).
[200] J. Fournier, L. Brossard, J. Y. Tilquin, R. Cote, J. P. Dodelet, D. Guay and H. Menard,
J. Electrochem. Soc., 143, 919 (1996).
[201] G. Kreysa and B. Hakansson, J. Electroanal. Chem, 201, 61 (1986).
[202] J. Divisek, P. Malinowski, J. Mergel and H. Schmitz, Int. J. Hydrogen Energy, 13, 141
(1988).
[203] J. Balej, Int. J. Hydrogen Energy, 10, 89 (1985).
[204] D. E. Hall, J. Electrochem. Soc., 132, C41-C48 (1985).
[205] R.B.Ferguson, NASA Research Center, NAS Report: 19700044586, (1969).
[206] J. Divisek and P. Malinowski, J. Electrochem. Soc., 133, 915 (1986).
[207] H. Wendt and H. Hofmann, Int. J. Hydrogen Energy, 10, 375 (1985).
[208] P. Schmittinger, in Electrochemical Hydrogen Technologies, H. Vendt, Editor, p. 261,
Elsevier, Amsterdam (1990).
[209] V. Utgikar and T. Thiesen, Int. J. Hydrogen Energy, 31, 939 (2006).
[210] N. Osada, H. Uchida and M. Watanabe, J. Electrochem. Soc., 153, A816-A820 (2006).
[211] J. Rifkin, The Hydrogen Economy, Tarcher, New York (2002).
192
[212] U. Bossel, in Intelec '05, p. 659, IEEE, Berlin (2005).
[213] U. Bossel, B. Eliasson, and G. Taylor, E08, EFCF, Lucerne (2005).
[214] I. Chorkendorff and J. W. Niemantsverdriet, Concept of Modern Catalysis, WileyVCH, Weinheim (2003).
[215] A. O. Isenberg, Solid State Ionics, 3-4, 431 (1981).
[216] W. Dönitz, E. Erdle, and R. Streicher, in Electrochemical Hydrogen Technologies, H.
Vendt, Editor, p. 213, Elsevier, Amsterdam (1990).
[217] S. H. Jensen, J. V. T. Høgh, R. Barfod and M. Mogensen, in Risø international energy
conference, S. L. Petersen and H. Larsen, Editors, p. 204, Risø National Laboratory,
Roskilde (2003).
[218] S. H. Jensen and M. Mogensen, in 19th. World Energy Congress, World Energy
Council, London (2004).
[219] H. Uchida, N. Osada and M. Watanabe, Electrochemical and Solid State Letters, 7,
A500-A502 (2004).
[220] S. H. Jensen, P. H. Larsen and M. Mogensen, Int. J. Hydrogen Energy, to be published
(2007).
[221] A. Hagen et al., in 6th European SOFC Forum, M. Mogensen, Editor, p. 930,
European Fuel Cell Forum, Lucerne (2004).
[222] S. H. Jensen, A. Hauch, I. Chorkendorff, M. Mogensen and T. Jacobsen, J.
Electrochem. Soc., -to be published (2007).
[223] A. Hauch, S. H. Jensen, S. Ramousse and M. Mogensen, J. Electrochem. Soc., 153,
A1741-A1747 (2006).
[224] M. Ni, M. K. H. Leung and D. Y. C. Leung, Chemical Engineering & Technology, 29,
636 (2006).
[225] M. Ni, M. K. H. Leung and D. Y. C. Leung, J. Power Sources, 163, 460 (2006).
[226] C. Mansilla, J. Sigurvinsson, A. Bontemps, A. Marechal and F. Werkoff, Energy, 32,
423 (2007).
[227] R. Hino, K. Haga, H. Aita and K. Sekita, Nuclear Engineering and Design, 233, 363
(2004).
[228] J. E. O'Brien, C. M. Stoots, J. S. Herring and J. J. Hartvigsen, J. Fuel Cell Science and
Technology, 3, 213 (2006).
[229] A. Hauch, S. H. Jensen and M. Mogensen, J. Electrochem. Soc., -accepted (2007).
[230] H. S. Hong, U. S. Chae, S. T. Choo and K. S. Lee, J. Power Sources, 149, 84 (2005).
[231] H. S. Hong, U. S. Chae, K. M. Park and S. T. Choo, Eco-Materials Processing &
Design Vi, 486-487, 662 (2005).
[232] W. S. Wang, Y. Y. Huang, S. W. Jung, J. M. Vohs and R. J. Gorte, J. Electrochem.
Soc., 153, A2066-A2070 (2006).
[233] B. Zhu, I. Albinsson, C. Andersson, K. Borsand, M. Nilsson and B. E. Mellander,
Electrochemistry Communications, 8, 495 (2006).
[234] T. Schober, Solid State Ionics, 162, 277 (2003).
[235] B. Yildiz and M. S. Kazimi, Int. J. Hydrogen Energy, 31, 77 (2006).
[236] B. Yildiz, K. J. Hohnholt and M. S. Kazimi, Nuclear Technology, 155, 1 (2006).
193
[237] E. Erdle, W. Dönitz, R. Schamm and A. Koch, Int. J. Hydrogen Energy, 17, 817
(1992).
[238]K. Eguchi, T. Hatagishi and H. Arai, Solid State Ionics, 86-8, 1245 (1996).
[239]A. Momma, T. Kato, Y. Kaga and S. Nagata, J. Ceramic Soc. Japan, 105, 369 (1997).
[240] J. S. Herring et al., in Second Information Exchange Meeting on Nuclear Production of
Hydrogen, p. 1, Argonne National Laboratory, Illinois (2007).
[241] J. E. O'Brien et al., in 12th International Conference on Nuclear Engineering, ASME,
Virginia (2004).
[242] M. Ni, M. K. H. Leung and D. Y. C. Leung, Chemical Engineering & Technology, 29,
636 (2006).
[243] S. M. Haile, Acta Materialia, 51, 5981 (2003).
[244] R. M. Ormerod, Chem. Soc. Rev., 32, 17 (2003).
[245] J. M. Ralph, A. C. Schoeler and M. Krumpelt, J. Mat. Sci., 36, 1161 (2001).
[246] S. C. Singhal, Solid State Ionics, 135, 305 (2000).
[247] K. K. Hansen, P. H. Larsen, Y. L. Liu, B. Kindl and M. Mogensen, in 5. European
SOFC forum, J. Huijsmans, Editor, p. 875, European Fuel Cell Forum, Oberrohrdorf
(2002).
[248] N. Christiansen et al., in 5. European SOFC forum, J. Huijsmans, Editor, p. 34,
European Fuel Cell Forum, Oberrohrdorf (2002).
[249] S. H. Jensen, A. Hauch, P. V. Hendriksen and M. Mogensen, J. Electrochem. Soc., to
be published (2007).
[250] M. Mogensen, P. V. Hendriksen and K. K. Hansen, in 5. European SOFC forum, J.
Huijsmans, Editor, p. 893, European Fuel Cell Forum, Oberrohrdorf (2002).
[251] NIST Chemistry WebBook, National Institute of Standards and Technology (2007).
http://webbook.nist.gov/chemistry/
[252] Orkustofnun, The National Energy Authority and the ministry of industry, Reykjavík
(2006). http://www.os.is/Apps/WebObjects/Orkustofnun.woa/wa/dp?id=920
[253] Statistical Yearbook of Norway, Statistics Norway (2006).
http://www.ssb.no/en/yearbook/tab/tab-250.html
[254] Fuel Cell Handbook V, Department of Energy, Morgantown (2000).
[255] Water Wise, ECO-TEC Inc. (2006). http://www.ecotec.com/pdf/Water%20Wise%200201.pdf#search=%22demineralized%20water%20cos
t%22
[256] Webshop, ScubaGear (2007). http://www.scubagear.dk/?id=1116e9da-719e-43e5871a-bc3ee46beae7&prodid=CYL10
[257] Prisliste, ICM Sikkerhedsmaterial (2007). http://www.icmas.dk/ICM%20priser%20010307.pdf
[258] N. Christiansen et al., in SOFC IX, S. C. Singhal and J. Mizusaki, Editors, p. 168,
Electrochem. Soc., Quebec City (2007).
[259] G. A. Mills, Fuel, 73, 1243 (1994).
194
[260] J. B. Hansen, in Handbook of Catalysis, G. Ertl, H. Knözinger and J. Weitkamp,
Editors, p. 1856, Wiley-VCH, New York (1997).
[261] M. H. Maack and J. B. Skulason, J. Cleaner Production, 14, 52 (2006).
[262] B. Arnason, T. I. Sigfusson and V. K. Jonsson, Int. J. Hydrogen Energy, 18, 915
(1993).
[263] A. H. Pedersen, Dansk Kemi, 82, 20 (2001).
[264] G.Astarita, D.W.Savage, and A.Bishio, Gas Treating with Chemical Solvents, John
Wiley & Sons, (1983).
[265] W.J.J.Huijgen and R.N.J.Comans, Carbon dioxide storage by mineral carbonation,
Energy Research Centre of The Netherlands, Editor, Cheltenham, United Kingdom,
(2005).
[266] B.Metz, O.Davidson, H.de Coninck, M.Loos, and L.Meyer, Carbon capture and
storage, Intergovernmental Panel on Climate Change (IPCC), Editor, Cambridge
University Press, Cambridge, United Kingdom and New York, NY, USA., (2005).
[267] C.W.Bale, P.Chartrand, S.A.Degterov, G.Eriksson, K.Hack, R.Ben Mahfoud,
J.Melançon, A.D.Pelton, and S.Petersen, FactSage Thermochemical Software and
Databases, CRCT - Center for Research in Computational Thermochemistry École
Polytechnique (Université de Montréal), Box 6079, Station Downtown, Montréal,
Québec, CANADA and GTT-Technologies, Kaiserstrasse 100, 52134 Herzogenrath,
GERMANY, (2002).
[268] I. M. Petrushina, N. J. Bjerrum, V. A. Bandur and L. N. Cleemann. Topics in Catalysis.
44(3)427-434 (2007)
[269] B. Kroposki, J. Levene, and K. Harrison, “Electrolysis: Information and Opportunities
for Electric Power Utilities”, NREL, 2006.
[270] S. Kristensen ”Elektrolyseanlæg i DONG Energy’s vision for Venzinproduktion”,
DTU, 2006
[271] “Special Transformers, Furnace and Rectifier Transformers”, ABB Brochure
[272] N. Eaton, G. Murison, and B. Speer, “Specifying rectifiers for electrochemical
applications,” in Conf. Rec. IEEE-IAS PCIC, 1997
[273] Maniscalco, Scaini, Veerkamp, “Specifying DC Chopper Systems for Electrochemical
Applications”, IEEE Transactions On Industry Applications, 2001
[274] Eltra/Elkraft review on flexible electrical load, “Pris elastisk elforbrug 2005”,
Energinet.dk, 2005
[275] Energinet.dk website, “udtræk af markedsdata”.
[276] ”Udbudsbetingelser for reguleringsreserver og systemtjenester, teknisk del”,
Energinet.dk, 2005
[277] “TF 5.9.1_vest, systemydelser og systemydelser for elproduktionsanlæg tilsluttet
kollektive net”, Energinet.dk, 2004
[278] “TF 5.9 Systemydelser”, Energinet.dk, 2004
195