HYDROGEN PRODUCTION BY ELECTROLYSIS

KNUT ANDREASSEN
Norsk Hydro Electrolysers AS
P.O. Box 44, N-3671 Notodden, Norway

Abstract

Industrial electrolysers are mostly using alkaline electrolyte and are operating at about
80 °C and 1 - 30 bar. Cells based on solid electrolytes, polymers or ceramics, are
developed, but have for the time being no commercial significance. The best available
technology of today may have an electric energy efficiency of about 80 %. Two main
types of electrolysers are in use, "tank cells" with electrodes connected in parallel and
"filterpress cells" with electrodes in series. In a plant the former electrolysers are
connected in series, the latter ones in parallel and eventually also in series. The hydrogen
production cost is under certain assumptions about 1,40 NOK/Nm3 or 0,40 NOK/k:Wh.

1. Introduction

1.1. ROUTES FOR HYDROGEN PRODUCTION

Based on various energy sources and hydrogen containing compounds as raw materials,
the different routes for the production of hydrogen are illustrated in Figure 1. Chemical,
thermochemical, electrical and photochemical conversion processes are used. Today
steam reforming/cracking of hydrocarbons and electrolysis of water are used in an
industrial large scale.
With a fossil hydrocarbon as the main energy source and raw material, and using a
chemical conversion process, 30- 50% of the hydrogen originates from the hydro-
carbon, the rest from water. Water is the main raw material in the other routes. Practi-
cally, they are based on electricity production and electrolytic water splitting. This
method is therefore the most direct and general one.

1.2. WATERELECTROLYSIS

The history of water electrolysis starts as early as the first industrial revolution. In the
year 1800, Carlisle and Nicholson demonstrated splitting of water by electrolysis.
91
T. 0. Saetre (ed.), Hydrogen Power: Theoretical and Engineering Solutions, 91-102.
@ 1998 Kluwer Academic Publishers.
92

In fact, it is the oldest electrochemical process known. However, nearly 100 years had
to pass away before it was introduced as a useful industrial way of producing hydrogen
and oxygen. In 1902 about 400 electrolysers were in operation. Norsk Hydro started its
first electrolysis plant in 1928, and electrolytic generated hydrogen was the main basis
for the company's fertilizer production for 50 years. The production reached a peak in
the sixties, amounting to about 100 000 Nm3/h.
Presently only 0,5% of the total hydrogen production is generated by water electro-
lysis. However, in the future when the use of hydrocarbons has to be restricted as a
result of limited resourCes and/or environmental pollution, this process may take the lead.

Chemical h
Fossil Energy~
----------- Conversion 1
Thermochem. .
~j
· ~Heat I...~,-----_,..·- - - - 1 Cycles \
· ~ ElectriCity 1 Electrolysis ~

Geothermal! . J Thermochem. \
Energy ~ Heatk::::::-------11 Cycles ~
IElectricity Electrolysis

I
Thermochem.
Nuclear Energy J-----1 Heat 1~----=:::::::::----------1 Cycles
------J Electricity Electrolysis f---~ HYDROGEN

I Thermochem.
Heat! ----~Cycles
~Electricity Electrolysis
Solar Energy Cbemical

Energy __ __ _ __ ~ Bioc~alr

I=·~
Water-.rl
Wind, Waves,
Ocean Currents
I
Ekrtricityl

Electricity Electrolysis I
Figure 1. Hydrogen Production Routes

2. Principles

2.1. GENERAL

An electrolysis cell is composed of two electronic conductors, electrodes, in contact with

an ionic conductor, electrolyte. When an electric direct current is passed through a cell
containing water, the water molecule may be split into hydrogen and oxygen.
 93

~0 + energy= ~ + 1/2 0 2. (1)

Water is a very poor ionic conductor, and the electrolyte has to contain also other
components to obtain a reasonable conductivity. These components should not undergo
any changes during the electrolysis.

2.2. ALKALINE ELECTROLYTE

The discovery of electrolytic water splitting was done in an acidic solution. Despite this
fact, in industrial plants an alkaline medium has been preferred up to now, mainly
because of lesser corrosion problemes.
In Figure 2 the principles of an electrolysis cell with a KOH-solution as electrolyte
are shown.
The basic reactions at the two electrodes are:
Cathode: 2 ~0 + 2 e· = ~ + 2 OR (2)
Anode: 2 OR = 1/2 0 2 + ~0 + 2 e· (3)
The net reaction is given above, reaction (1).
Between the electrodes there is a separator, a diaphragm or a membrane, to keep the
produced gases apart. Hydroxylium ions migrate through the separator during the
electrolysis process. The process water is generally fed to circulating electrolyte.
KOH is the most used alkaline component because of good conductivity. The con-
centration may be 25 - 30 %by weight. The working temperature has then to be lower
than 100 °C, usually about 80 °C. The pressure could be from 1 to 30 bar.
The cathode and the anode must be corrosion resistant in the electrolyte at reducing,
respectively oxidizing potentials, be good electronic conductors and have structural
integrity. They should catalyze evolution of the gasses as effectively as possible.

KOH
cathode diaphragm anode

Figure 2. Alkaline Electrolyte

2.3. ACID ELECTROLYTE

Today acid solutions are not used in water electrolysis practice. However, acid mem-
branes have been introduced as interesting electrolytes, and the technology is under
continous development. The principles are shown in Figure 3.
94

In this case the basic reactions at the electrodes are:

Cathode: 2 H' + 2 e- = ~ (4)
Anode: Hp = 1/2 0 2 + 2 H' + 2 e- (5)
The net reaction is of course the same as before, reaction (I) above.
Protons are migrating through the membrane during electrolysis. The membrane
does not only serve as a solid electrolyte, but also as gas separator. Water may only be
fed to the anode chamber of the cell.
The membrane is a so called proton exchanger. It is mostly made of polymer
materials_ Perfluorocarbon sulfonates are commonly used. The upper temperature limit
for these is 125 - 150 °C, but in general liquid water and an operation temperature of
about 80 °C is used.
Also ceramic proton conducting materials could be used as membranes. The ope-
rating temperature is then in a range where water is present as steam.
The electrodes are in these cells always placed in direct contact with the membrane.
They could be thin layers of catalyst materials.
This electrolysis system allows high current densities at high cell efficiency. The
drawback is high investment cost because of expensive membrane and electrode
materials.
- +
r----<J <>----, .-----o- +
<>---,

4 4 4

3 3 2

Figure 3. Acid Membrane Electrolyte Figure 4. Oxygen Ion Conducting Electrolyte

l. Cathode 3. Solid Electrolyte
2. Anode 4. Current Collector

2.4. OXYGEN ION CONDUCTING ELECTROLYTE

Electrolysis cells operating at high temperature using proton conducting ceramics as

electrolyte are only tested in small laboratory scale. Much more development work is
performed concerning oxygen ion conducting solid electrolytes. The principles are
shown in Figure 4.
 95

The basic reactions at the electrodes are:

Cathode: HzO + 2 e· = H2 + 0 2- (6)
Anode: 0 2' = 112 02 + 2 e· (7)
The net reaction is again reaction ( 1) above.
In this case oxygen ions are migrating through the solid electrolyte, serving also as a
gas separator. Water is fed to the cathode chamber of the cell.
The electrolyte is made of ceramic materials. The operating temperature has to be
high, 700 - 1 000 °C, because only in this temperature range known ceramics have a
satisfactory ionic conductivity.
The cells could operate at a high current density and have high efficiency. The draw-
back is that all construction materials have to be ceramics because of the high tempera-
ture used. This creates design problems, high investment and maintenance costs.

3. Energy Consumption

3.1. THERMODYNAMICS

The water splitting reaction is as stated before:

HzO = Hz + 112 0 2 (8)
Under standard conditions (component activity: 1, temperature: 273,15 K, presure:
1 bar) the Helmholtz energy (reaction enthalpy) based on liquid water is
Alf273 L = 285,9 kJ/mol.
This corresponds to the so called highe~ heating value of hydrogen. When the splitting is
performed by electrolysis, the main part of the total reaction energy has to be supplied as
electric energy, corresponding to the Gibbs energy. Under standard conditions we have:
AG0273 L = 237,2 kJ/mol.
The well known relation between Helnmoltz and Gibbs energy is:
AH = AG + T AS (9)
The energy T · AS could thermodynamically be covered by thermal energy.
The cell voltages corresponding to the two thermodynamic energies are:
The thermoneutral cell voltage: Utn = AH/(2 F) (10)
The reversible cell voltage: Urev = AG/(2 F) (11)
AH and AG and then also Uth and Urev are functions of electrolyte composition,
temperature and pressure. In Figure 5 the relationships to temperature are shown. The
brake of the curves at 100 °C is of course due to the transition from liquid water to
steam. It corresponds to the evaporation energy.
The energy necessary to heat the water from e.g. 25 °C to the reaction temperature
T, to rise the pressure from 1 bar to the reaction pressure p, to evaporate water and to
compensate for heat losses could thermodynamically be covered by thermal energy. In
industrial electrolysis a greater part is however covered by electrical energy, as shown
below.
96

The electrical energy efficiency of the electrolysis process could be defined by the
equation:
e = Utn.m.r. I U (12)
where U is the actual cell voltage.

..
...
••
Figure 5. Voltage I Temperature Relationships Figure 6. Voltage I Current Density Relationships

3.2. CELL VOLTAGE

Water electrolysis under technical conditions demands a cell voltage which is noticeable
higher than the reversible value. The voltage has to overcome the electrical resistance in
the electrodes, in the electrolyte between the electrodes and in the separator. In addition
come also the overvoltages at the electrodes, which are kinetically determined. The
actual cell voltage is then given by the relation:
U = U,.., + YJ, + ll \ + j R. ( I3 )
where llc is the overvoltage at the cathode, in this case the so called hydrogen over-
voltage, llA is the overvoltage at the anode, the oxygen overvoltage, j is the current
density and R" the sum of the electrical resistances, based upon the active electrode area.
Figure 6 shows an example of the relations between voltages and current density in
conventional alkaline electrolysis.
The conditions could also be illustrated as shown in Figure 7. Examples of real
values are given in Table I, refering to an industrial electrolyser using alkaline electrolyte
at common conditions (25% KOH, 80 °C, I bar).
The electrical resistance in the electrodes is determined by the material and design of
the electrodes and the temperature. The resistance in the interface electrode/electrolyte
is a function of for one thing the wetting conditions. The resistance in the electrolyte
layer between the electrodes is determined by the resistivity of the electrolyte (function
of composition and temperature) and the electrode gap. The resistance in the diaphragm
is dependent of the material and the way it is fabricated (e.g. woven). The overvoltages
are determined by the activity of the electrode surfaces, the electrolyte composition and
temperature.
 97

TABLE 1. Figure 7. Cell Voltage Elements

Ohmic Resistances and Cell Voltage Elements ....-...-: /
~u. ~
Electrodes: R"A+R"c = 2,0·10.,;Qcm2 / /
/ u., /
Electr. Surfaces: RAl+ R"c,= 31,2 · 10.,; " /
Electrolyte: R"E = 41,4 · 10-6 " / u.. /
/

Separator: R"0 = 36,0 · 10.,; / .... /

Sum R" = 110,6 · 10-6 Q cm2
/-"-•---- ./
~ "~ - - - - -1);- . /
/

Rev. Cell Voltage: U,.. = UAt + Uc 1 = 1,189 V 0

/- ""
\)..._/
/ ,/
Overv. (j=0,20 Ncm2): u. = llc + 11A = 0,325 "
/
Ohm.v.(j=0,20 Ncm2): U0 = j · R" = 0,221"
.~,.,

/
Sum: = 1,735 V
;,
We see that the reversible cell voltage in the example makes about 68 % of the actual
cell voltage, while the overvoltages and the ohmic voltage drop makes about 19 % and
about 13% of the total value respectively.

3.3. ENERGY CONSUMPTION

The energy consumed during hydrogen production is given by the following relationship:
Em = 2 · F · Pm I (3600 · Mm) · U I c (14)
where F is the Faraday constant, c the current efficiency, Pm the density and Mm the
molar mass of hydrogen. Puting in the values we have for the example above:
Em = 2 · 26,80 · 0,08988 I 2,016 · U I c = 2,39 · 1,735 I 0,98 = 4,23 kWh I m3

4. Cells, Electrolysers and Plants

4.1. GENERAL

The description in this chapter will be limited to technology based on alkaline electro-
lytes, since for the time beeing these are the only ones used industrially, generally as
KOH-solutions.

4.2. CELLS

As an example of an industrial cell configuration Figure 8 shows the elements ofthe

Norsk Hydro design. It is made of two electrodes, two gaskets and a frame carrying the
separator or diaphragm. At the bottom of the cell there are openings which form the
electrolyte supply and distribution ducts. Hydrogen and oxygen respectively flow to
openings at the top of the cell, forming ducts for gas and electrolyte removal. The cell is
operating with 25 % KOH at 80 °C and I bar.
98

Figure 8. Norsk Hydro Cell

The gases evolving at the electrodes escape as bubbles. These are whirling around in
the electrolyte and block the current paths between the electrodes. The gas/electrolyte
volume ratio should therefore be low. On the other hand the gap between the electrodes
should be small to minimise the ohmic resistance. Both are obtained in this case by using
perforated so called fore-electrodes. Even though the gas evolution principally takes
place on the front side of the fore-electrodes, the greater part of the gas bubles are rising
on the back side. In some other cell designs a zero gap configuration is used. The fore-
electrodes could be made of porous or extended metal sheets.
The diaphragm in the example is a woven cloth of a polymer material fixed in a
frame. To act as an efficient gas separator with low electrical resistance, the cloth must
be wetted by the electrolyte so that the pores are filled up. The frame and the electrodes
are made of nickel plated steel. The fore-electrodes are also activated by catalytic
coatings to lower the overvoltages.

4.3. ELECTROL YSERS

Industrial electrolysers are composed of many cells or electrode pairs. These may be
connected in parallel or in series. Figure 9 shows the principles of a parallel connection.
Each individual electrode is fed by current from outside and has only one polarity; it is
monopolar, i.e. is either cathode or anode. The principles of a series connection are
shown in Figure 10. Only the end electrodes are fed by current and are monopolar. The
electrodes in between are bipolar, i.e. cathode on one side and anode on the other.
Electrolysers having only monopolar electrodes are often called "tank cells" while
electrolysers with bipolar electrodes are called "filterpress cells" due to the appearance.
An electrolysis plant have generally several electrolysers, connected in series or in
series and parallel. The alternatives are shown in the Figures 11 and 12 respectively.
 99

Cll 0 I® 0 0 0 fll ~

J.: l. l.: J, J.: J, J.: J, J.: J,

L~W~fJWLfJWL L~W L WL~WL~

J, J.: J, l. J, l,
'' '' '' '' '' ''
' '' '' '' '' '''
'' ' ' ' '
L . . _ - - - - - - 0 • • x "e•U G - - - - - - . . . J

Figure 9. Parallel Connection of Cells Figure 10. Series Connection of Cells

e
0~------------~ e
01 ,--, ,--,
'---1....J-1.__------------

Figure 11. Series Connection of Electrolysers Figure 12. Parallel and Series Connection of Electrolysers

Each of the two electrolyser designs has advantages and disadvantages. The tank
cell is simple and sturdy, while the filterpress cell for one thing has a lot of gaskets. The
outer leeds to the tank cell cause voltage loss. The filterpress cell has channels for e.g.
electrolyte supply, forming shunts to the cell stack and then causing current loss. The
tank cell has a low voltage and is contingent upon a high current to give a reasonable
production. The filterpress cell has a high voltage and operates with a low current.
However, both are settled in a plant having several electrolysers.
As an example of a filterpress cell Figure 13 shows the Norsk Hydro design. The
cell package is clamped together between rigid steel frames. The bus bars are connected
to the front and rear ends.

Figure 13. Norsk Hydro Electrolyser

100

4.4. PLANTS

The equipment around the electrolysers in a plant may vary from place to place and from
provider to provider, even though the principles are not very different. As an illustration
one ofNorsk Hydro's systems is shown in Figure 14.

Figure 14. Electrolyser Plant

The produced hydrogen and oxygen gases together with some electrolyte flow
through the removal ducts to the gas/electrolyte separators. The hydrogen is purified,
cooled and dried and eventually compressed. Excess heat from the electrolyser is
removed from the circulating electrolyte in the cooler. The feed water is purified in an
ion exchanger and added to the electrolyte previous to its entry in the electrolyser' s
supply duct.

5. Economy

5. 1. GENERAL

The economy of a hydrogen producing electrolysis plant is as usual dependent of the

investment and the maintenance and operating costs, all in relation to the production
rate.

5.2. COSTDATA

In Table 2 are shown some characteristic data for a plant having a producing capacity of
4 000 Nm3 H 2 I h.
 101

TABLE 2. Cost estimates

Plant Capacity 4 000 Nm3 Hth

Total investment cost: NOK 90 000 000

Annual capital cost: NOK 10 500 000
Annual maintenance and operating cost: NOK 5000000
Electricity cost: NOK 31 500 000
Total annual cost: NOK 47 000 000
Specific cost: NOK 1,42 pr. Nm3 ~
NOK 0,40 pr kWh (HHV of~)

The cost estimates are based on specific conditions, requirements and limitations.
The most decisive are:
* The plant is located on an autonomous site equiped with its own administration,
maintenance, plant control facilities and infrastructure.
* High voltage power supply and utilities such as treated and untreated water and a
disposal system are readily available at reasonable cost.
* Hydrogen should be delivered at 30 bar at the plant battery limit.
* Oxygen could not be commercially used and is vented to air in plant area.
* Waste heat could not be commercialised.
* The economical lifetime of the plant is set to 15 years and the interest rate to 8 %.
* Hydrogen is produced in 8 300 hI year, the availability of the the plant is about 95%.
* Reactivation of the electrodes is done every 7 years.
* The total specific energy consumption is 4, 75 kWh I Nm3 ~- This include the
consumption in the electrolysers, the bus bars, the transformers, the compressors and
other auxiliary equipment
* The electricity price is stipulated to NOK 0,20 pr. kWh.
The table shows that in this case the capital cost is about 22 % of the total pro-
duction cost, the maintenance and operating costs are about 11 % and the electricity cost
about 67%.

5.3. REDUCTION OF COST ELEMENTS

In the example above the electricity cost is the dominating element in the total cost
picture. It is then to emphasize that nearly 60 % of the electricity consumption is due to
the Gibbs energy. Reduction of this element can practically be obtained only by rising
the working temperature. The main argument for developing high temperature
electrolysis processes is just to attack the thermodynamic part of the electricity
consumption.
However, in the near future the research and development work seems to be centred
in electrolysis at midium and low temperatures. We will here only discuss the latter, and
then limit ourselves to electrolysis in alkaline solution.
102

5.3 .1. J<..nergy Consumption I Energy Cost

The energy consumption may be reduced by lowering the ohmic voltage drop, primarily
in the cells themselves. The following possibilities should be valued:
* Amend the electrolysis conditions ( lower the current density, rise the pressure, rise
the temperature, change the electrolyte concentration )
* Reduce the resistance in the electrolyte layer between the electrodes ( diminish the
gap, reduce the gas/electrolyte volum ratio, secure the electrolyte circulation )
* Reduce the resistance in the diaphragm ( choose another diaphragm type )
* Reduce the resistance in the electrodes ( the fore an back electrodes )
The energy consumption may also be reduced by lowering/stabilizing the
overvoltages. The following possibilities should be valued:
* Amend the electrolysis conditions ( lower the current density, rise the temperature )
* Improve/stabilize the activity of the electrocatalysts I active coatings
5.3 .2. Investment I Capital Cost
The specific investment may be reduced by reducing the production costs of cells,
electrolysers and total plant. The following possibilities should be valued:
* Simpler cell and electrolyser design
* Cheaper construction materials
* More efficient manufacture
* More efficient installation
* Simpler design of the total plant
Some of the possibilities are more or less coupled together, especially concerning the
cells, which make a greater part of the cost. By using other construction materials it may
be practicable to change both the design and the manufacture.
The specific investment may also be reduced by increasing the production rate of
cells and electrolysers, i.e. the current. The following possibilities may be valued:
* Enlarge the cell dimensions
* Rise the current density
Increased gas production in the cells may be contingent upon a change of the electrolyser
and plant design.

5.3.3. Economical Optimisation

Many of the efforts pointing at reducing the energy cost may also affect the investment,
and vice versa. As usual an economical optimisation has to be done.

5.4. POTENTIAL OF IMPROVEMENTS

As concerns electrolysis in alkaline electrolyte it should be possible to reduce the

investment I capital cost by say 30 % and the energy consumption by 10 - 15 %. Based
on the figures presented above the hydrogen production cost then could be lowered by
about 15%. The specific cost is then reduced to 1,20 NOK/Nm3 or 0,34 NOK/k:Wh.