Elyses em

ALCAN ALESA ENGINEERING LTD.
Electrolysis Seminar
March, 2007
Chippis, Switzerland
Max-Högger-Strasse 6
CH - 8048 Zürich
Switzerland
Telephone: +41 44 435 33 33
Facsimile: +41 44 435 06 65
E-Mail: [email protected]
Website: www.alcanalesa.ch
Authors: Topics:
Peter Entner Theory of Aluminum Smelting,
Handbook, ElyseSem Acrobat-File, ElysePrg PC-Program.
e-mail: [email protected].
website: www.peter-entner.com.
Ulrich Heinzmann Practice of Aluminum Smelting.

e-mail: [email protected].
All copyrights and other rights regarding the contents and enclosures of the Electrolysis
Seminar Book are with ALCAN ALESA ENGINEERING LTD. Without the express
permission of the management of this company it is forbidden to copy the Electrolysis
Seminar Book either in whole or in part or make it accessible to third parties.
ALESA Electrolysis Seminar
Contents
Contents Electrolysis Seminar
Theory of Aluminum Smelting

1. Principles of the Hall-Héroult-Process
2. Mass Balance
3. Properties of the Electrolyte
4. Pot Voltage
5. Energy Balance
6. Current Efficiency
7. Thermal Effects
8. Magnetic Effects
Practice of Aluminum Smelting
9. Raw Materials
10. Cathode Lining and Pot Startup
11. Process Control
12. Pot Operation
13. Troubleshooting in Pot Operation
14. Measurements
15. Autopsies
16. Emissions from Aluminum Electrolysis
17. How to Use ElyseSem

18. How to Use ElysePrg

Max-Högger-Strasse 6 CH-8048 Zürich, Switzerland
Tel: +41 44 435 3333, Fax: +41 44 435 0665, E-Mail: [email protected]
Contents

1. Principles of the
Hall-Héroult-Process
1.1. Summary ...................................................................... 1-3
1.2. Aluminum Production by Electrolysis....................... 1-3
1.3. Industrial Production of Aluminum ........................... 1-4
1.4. Alternative Production Processes........................... 1-12
1.5. Inert Anodes .............................................................. 1-14
1.6. Wettable Cathodes .................................................... 1-18
1.7. Literature.................................................................... 1-21

1-1
ALCAN ALESA ENGINEERING Ltd.

1-2
1. Principles of the Hall-Héroult Process
1.1. Summary
The principles of the Hall-Héroult process to produce aluminum

in an electrolytic cell are shown. Alumina, an aluminum oxide
(Al203), is dissolved in molten cryolite (Na3AlF6). It is decom-
posed electrolytically between carbon and aluminum electrodes
at about 950° C to aluminum and oxygen. The carbon anode is
continuously consumed by reacting with the oxygen to give
carbon dioxide (CO2).
The typical features of the electrolytic cells in an aluminum
plant are described. Technical cells are operated with a current
intensity of 100 to 300 kA. They are equipped with devices to
operate the cells in a highly automatic i.e. computer controlled
way. The waste gases of the process are collected and thor-
oughly cleaned (scrubbed) before being exhausted to the at-
mosphere.
The discussion of alternative methods to produce aluminum in
technical quantities mentions the Direct Carbothermic Reduc-
tion Process and the Alcoa Smelting Process which were devel-
oped to replace the classical Hall-Héroult process. Technical
problems which could not be solved stopped however these
projects.
Inert anodes must sustain the chemical attack of the anodic
oxygen and of the electrolyte at electrolysis temperature. Sev-
eral candidate materials were investigated however without
convincing success so far.
Using materials which are wetted by aluminum as cathode
would avoid several problems essentially the magnetic effects
and the chemical reactions of the carbon lining. Several patents
and publications propose materials and arrangements to re-
place the aluminum metal pad as cathode.
1.2. Aluminum Production by Electrolysis
In 1886 Hall of the USA and Héroult of France invented simultaneously

and independently of each other the process to produce aluminum by
electrolysis. Alumina, an oxide of aluminum (Al203), is dissolved in molten
cryolite (Na3AlF6) and decomposed electrolytically to give liquid aluminum.
The anode of the electrolytic cell is made of carbon and the pool of already
produced aluminum acts as cathode. The oxygen of the alumina is dis-

1-3
charged at the anode where it reacts with the carbon anode to produce
carbon dioxide (CO2).
Figure 1-1: Schema of a Hall-Héroult electrolysis cell for the production of aluminum.
In Figure 1-1 such a Hall-Héroult electrolysis cell is schematically shown.

The cryolite electrolyte and the liquid aluminum are contained in a steel
shell which is lined with carbon blocks and thermal insulation material.
Electrical energy is used to reduce electrolytically aluminum oxide and to
keep the electrolyte at a temperature of about 950° C. During aluminum
production the chemical reactions consume continuously alumina and an-
odes which must be added respectively replaced to the electrolytic cell.
Aluminum and anode gases (in essence carbon dioxide and carbon mon-
oxide) are produced and removed from the cell.
1.3. Industrial Production of Aluminum
In Figure 1-2 you see a photograph of an industrial electrolysis cell

(EPT18, the former 180 kA pot of Alusuisse) arranged side by side e.g. the
long sides of the pots are facing each other.
Fig
Figure 1-2: Photograph of an industrial electrolysis cell.

1-4
Some technical and operating figures of this pot type are given in Ta-
ble 1-1.
Table 1-1: Technical and Operational Figures of Alusuisse's EPT18 Pot (Lit. 1-1).
Figure 1-3 shows the corresponding schematic cross section of such a

technical pot.
Figure 1-3: Cross section of an industrial electrolysis cell.

1-5
Anodes
The carbon anodes are attached to the iron yoke and aluminum rod with
the anode clamp to the anode beam. After an anode is consumed the an-
ode clamp is opened and the consumed anode is replaced by a new one.
A hoist motor changes the position of the anode beam to control in this
way the distance between the electrolytic active bottom side of the anode
and the upper surface of the metal, the interpolar distance.
Pot Shell, Lining, Bath and Metal Pad

The steel pot shell contains the liquid phases namely the electrolyte and
the molten aluminum. The pot shell is lined with carbon bottom blocks and
carbon side blocks. In addition thermal insulating bricks of the pot lining
control the thermal loss through the bottom and side walls. The electrolyte
freezes on the pot sides to a layer, the side ledge. This layer protects the
side and bottom carbon blocks against corrosion and erosion by the mol-
ten electrolyte and moving metal. Electrolytic bath freezes also on its up-
per surface to a solid top crust. On the top crust a layer of alumina or
crushed solid bath is piled up. This layer controls the thermal loss of the
upper surface of the electrolytic pot. It protects also the carbon anodes
against the reaction with the ambient air. The steel shell is the container
for the pot lining, the electrolytic bath and the produced aluminum. It is
surrounded with reinforcing girders to resist the forces of the thermal and
chemical expansion of the lining materials.
Electric Current Flow

On its way through the electrolytic pot the electric current is flowing
through the risers to the anode beam, then to the anodes, crosses the in-
terpolar distance and is conducted by the collector bars and the aluminum
bus bars to the following electrolytic pot. These busbars transport the elec-
tric current from one electrolytic pot to the next one either around the pot
shell and/or under it.
Hooding
The top side of the pot is closed by a hooding to collect the waste gases.
The pot gas exhaust conducts these gases to a scrubbing unit. The scrub-
ber mixes the waste gases with alumina to adsorb e.g. remove the gase-
ous and particulate fluoride components before the waste gas is ex-
hausted to the atmosphere.
Alumina
Since the electrolytic process consumes alumina at a nearly constant rate
alumina must be added to the electrolytic bath. In modern pots automatic
transport and feeding systems have replaced the vehicles which went from
one pot to the other, brought alumina from the silos to the pots and
crushed the top crust to push alumina into the bath. In patents (Lit. 1-2,
1-3) and the literature several systems are described to transport alumina
from the plant silo to the pots and to add a known amount of alumina to
the electrolyte.

1-6
The tubes which transport alumina are worn by the moving material and
on the other hand alumina tends to stick together in blind angles or block
the tubing if the speed is too small. To overcome these problems the con-
ventional fluidized bed was modified especially to decrease the amount of
compressed air to transport alumina.
Dense Phase Conveying System (Lit. 1-4)

Two parallel pipes are transporting alumina (Figure 1-4) and com-
pressed air. The compressed air enters into the alumina pipe through
openings which are filled with porous material. Plates release the air pres-
sure in such a way that about the same amount of air enters along the
alumina transporting pipe. The result is a smooth motion of alumina.
Figure 1-4: Principle of the dense phase conveying system.
Hyperdense System (Figure 1-5)

This system uses the capacity of alumina to become fluidized under low
pressure. When alumina is fed to the bath and the alumina level de-
creases in the pot hopper pressure is released. Alumina is fluidized locally
to fill the void which moves along the tube to the silo. Because of using low
pressure the hyperdense system is not able to transport alumina uphill, a
disadvantage for the modernization of old pot lines.
Figure 1-5: Principle of the hyperdense system.
Point Feeder (Figure 1-6)

1-7
The automatic closed transport system conveys alumina from the plant silo
to alumina hoppers which are placed between the anode beams.
Figure 1-6: Point feeder with metering device.
To design such a transport system one has to take especially into account
the abrasive power of alumina and its tendency to plug narrow tubes. At
the pots several pneumatic air cylinders activate several crust breakers to
keep holes open in the top crust. A dosing system adds alumina in defined
quantities to the electrolyte where it is dissolved. The process control sys-
tem activates the point feeders (crust breakers and dosing devices) ac-
cording to a sophisticated algorithm to keep the alumina content of the
electrolyte at the target value.
A critical component of the system is the dosing device which measures a
defined amount of alumina in a reproducible way.
High Current Load Pots

The aluminum industry has invested much development work to improve
the performance of the electrolytic pot especially to increase the current
load without losing the good production figures.
The following photographs show examples of such high amperage pots
and tables list the corresponding key parameters:

1-8
Table 1-2: Cell Performance of the Dubai Aluminium Co. Pot at 190 kA (Lit. 1-5):
Figure 1-7: The 200 kA Pot of Dubai Aluminium Co. Ltd.

1-9
Table 1-3: Operating Figures of the Pechiney 300 kA Pot (Lit. 1-6):
Figure 1-8: The 300 kA Pot of Pechiney.

1-10
Table 1-4: Operating Figures of the 300 kA Pot of VAW (Lit. 1-7):
Figure 1-9: The 300 kA Pot of VAW.

1-11
1.4. Alternative Production Processes
Since the technical introduction of the classical Hall-Héroult-Process and

especially in the eighties due to the energy crisis great efforts were made
to replace or to improve the classical process (Lit. 1-8).
Direct Carbothermic Reduction Process

Iron is produced by direct reduction of iron oxide with carbon in a blast
furnace. This process cannot be used to produce aluminum since alumina
reacts with carbon to carbides. Alcoa therefore studied the feasibility to
produce aluminum by direct carbothermic reduction of alumina silicon ores
(Lit. 1-9). The concept was to reduce a mixture of aluminum oxide, silicon
oxide by coke in a Blast Furnace. The combustion heat of burning a frac-
tion of the coke should keep the reaction zone in the blast furnace at a
temperature over 2000 °C.
Figure 1-10: Principle of the Direct Carbothermic Reduction Process of Alcoa. Coke re-
duces bauxite and clay in a blast furnace. The aluminum-silicon-iron alloy is
split up by fractional crystallization. In the last step a membrane cell pro-
duces pure aluminum and silicon.
A crystallizer splits the alloy up into its components by fractional crystalli-

zation and a diaphragm transport cell purifies the aluminum by electrolysis
(Figure 1-10).
Experience showed however that a totally combustion heated process was
not feasible(Lit. 1-10). The project failed because electric energy was nec-
essary to produce the aluminum silicon alloy (Lit. 1-11).

1-12
Alcoa Smelting Process

One of the other projects to replace the Hall-Héroult-Process was the Al-
coa Smelting Process (Figure 1-11).
The first step produces AlCl3 by Carbo-Chlorination e.g. the reaction of
alumina with carbon and chlorine:
Al2O3 + 3 C + 3Cl2 = 2AlCl3 + 3 CO2 (1–1)

2 2
Figure 1-11: Schema of the Alcoa Aluminum-Chloride Process. AlCl3 is produced by re-
acting alumina with carbon and chlorine. The AlCl3 is electrolysed in a multi-
polar cell and the chlorine is recycled into the process.
The AlCl3 is then decomposed to aluminum and chlorine by electrolysis:
2AlCl3 = 2Al + 3Cl2 (1–2)

Before this however can be done the produced aluminum chloride must be
treated chemically e.g. purified thoroughly from byproducts like hydrox-
ides, oxides and chlorinated hydrocarbons. Chlorinated hydrocarbons are
severe environmental hazards. Oxygen containing compounds deteriorate
the graphite electrodes of the electrolysis cell.
The electrolysis cell contains several bipolar electrodes (Figure 1-12). The
electrolysis current passes the electrolysis cell form the top (positive an-
ode) to the bottom (negative cathode). The intermediate electrodes are
cathodes at the top where aluminum is produced and anodes at the bot-
tom where chlorine is formed. Since the anode to cathode distance is kept
at a minimum great effort had to be made to design the surfaces of the
electrodes in such a way that the electrolysis products are drained away
without reacting with each other.
The chlorine produced during electrolysis is collected, purified and recy-
cled into the chlorination process.

1-13
Figure 1-12: Aluminum Chloride Electrolysis Cell. Several bipolar electrodes decompose
the molten aluminum chloride into aluminum and chlorine.
Alcoa has constructed a plant in Palestine, Texas which produced

15000 t/y using the Alcoa Smelting Process. After large efforts also this
project was stopped especially because of the problems to handle the
chemistry of the chlorination process.
1.5. Inert Anodes
It has been a subject of extensive research since the invention of the Hall-
Héroult process to replace the carbon anodes which are consumed during
the electrolysis process by anodes which are made of inert material. The
work on inert anodes is reviewed in several articles, at the time being
Lit. 1-12 is the last one. This publication leads the interested reader to the
other literature.
The overall reaction when employing inert anodes is
2Al2 O 3 = 4Al + 3O 2 (1–3)
e.g. oxygen is evolving at the anode. Although the absolute value of re-
versible potential for an oxygen evolving anode is approximately 1V higher
than that of a carbon anode, a lower oxygen overvoltage and lower ohmic
losses of non-consumable anodes could compensate.

1-14
Figure 1-13: Principle of an Inert Anode. The anode body consisting of refractory com-
pounds is covered with a layer of inert anode material. Doped tin oxide or
nickel iron oxides are examples of inert anode materials.
Figure 1-13 shows the principles of using an inert anode. The surface of
the anode body is covered with the inert anode material mostly oxides
which conduct the electric current. Depending on the surface material the
body consists of a refractory material which is able to sustain the high
temperatures in the electrolytic pot. Also the electric connection between
the anode material and the bus bar system must be studied.
The potential advantages of a system with inert anodes are:
• savings in operating costs: eliminating carbon anodes replacement and provid-

ing tighter control of the anode-cathode distance.
• anode product is oxygen: contribute to reducing greenhouse gas emissions
associated with aluminum production. Value of re-
claimable oxygen might be worth 3% of the produc-
tion cost of crude aluminum.
• dimensional stability: more compact and efficient cell design, such as bi-
polar configuration and reduced thermal losses.
The severe operating environment and the requirement of good electronic

conductivity limit the number of suitable materials for inert anodes. The
main requirements that materials for inert anodes should fulfill are as fol-
lows:
• resistant to chemical and electrochemical attack by the highly corrosive cryolite-based

electrolyte, which contains dissolved aluminum,
• stable with respect to the evolution of pure nascent oxygen,
• high electrical conductivity and low contact resistance to external current conductors,
• thermal stability (thermal shock and ductility) up to 1000 °C,
• relatively low cost, readily available,
• easy to fabricate into large finished shapes,
• low overvoltage for discharge of oxide ions and high overvoltage for fluorine dis-
charge,
• high electronic conductance,
• stable electrical connection to bus,
• adequate mechanical strength.
While a vast array of materials has already been patented for inert anode
application, none is yet in commercial use. Most recent research has con-
centrated on metals, cermets, oxide ceramics and cerium oxide coatings.

1-15
Metal Anodes
Metals and metallic alloys have excellent electrical conductivity and have
therefore been proposed as anode materials for aluminum electrolysis.
However, apart from some Noble Metals, the metals are attacked by oxy-
gen produced at the anode. For the aluminum reduction cell the anode can
comprise an alloy of Copper, Nickel and Iron or combinations of them,
upon which forms a protective oxide layer that under cell operating condi-
tions is thick enough to prevent chemical attack of the underlying metal,
yet thin enough to allow electronic current to pass without a substantial in-
crease in cell voltage. Tungsten, Nickel and Stainless Steel anodes
have also been investigated. However, these metals have proved insuffi-
ciently resistant to anodic attack and therefore unsuitable as inert anodes
in the production of aluminum by the Hall-Héroult process. It is important
to form a dense oxide layer on the total anode surface, otherwise the
metal anode suffers direct corrosion by the electrolyte. This attack is ag-
gravated when the anode is polarized and not passivated, and can result
in direct electrostripping of the metal components of the anode. The tech-
nical viability of metals as anode materials is still under investigation.
Cermets
Cermets are a combination of ceramics and metals. Cermets developd so
far are generally composed of primarily Nickel Oxide (NiO) and Nickel
Ferrite (NiFe204) which contain a copper-based metal phase. The oxide
matrix provides a corrosion resistant network to contain the electrically
conductive copper-based metal phase. Figure 1-14 (Lit. 1-13) shows
schematically an experimental anode arrangement using cermet material.
Figure 1-14: Principle of an Inert Anode made of a Nickel-Ferrite Cermet Material.
While polarized, the metal phase develops a coating which helps to protect
the cermet from corrosion. The corrosion resistance of such cermet an-
odes has been determined recently in cryolite-based high temperature and
cryolite-potassium fluoride-based low temperature systems. Corrosion in

1-16
the high temperature system is 2 to 3 times higher than that in the low
temperature system. A novel technique using electroless copper plating
instead of mechanical mixing with copper has produced cermets with bet-
ter electrical conductivity. When tested in aluminum reduction cells, how-
ever, the cermet anode showed inadequate corrosion resistance, particu-
larly at low alumina concentrations and high current densities.
Full-size anodes with advanced or grooved design have been manufac-
tured to compare fabrication by isostatic pressing, slip casting and injec-
tion molding respectively. Of the three approaches, isostatic pressing pro-
duced an anode with dimensions nearest to the target specifications, and
without serious macroscopic flaws. This approach appears the most prom-
ising for making advanced anodes for aluminum smelting. However, to op-
timize the physical properties and microstructure of the anode will require
substantial further work.
Oxides
Ceramic oxides all have a certain degree of solubility in molten cryolite and
generally have a high electric resistivity. Among several materials studied,
Tin Oxide (SnO2) has proved to be the most promising candidate for
manufacturing inert anodes. Several laboratory studies have been made
namely on thermal shock resistance and on the amount and forms of SnO2
solubility in cryolite-alumina melts. Different doping materials like the ox-
ides of copper, iron, manganese have been used in preparing SnO2 semi-
conductor ceramics and the properties of these materials determined. Cell
tests using anodes with those compositions demonstrated fairly good re-
sults concerning current efficiencies, corrosion rates and anode surface
behavior.
Cerium Oxide Coating

In 1982 it was found that cerium oxide coatings could be anodically depos-
ited from a molten cryolite bath onto a variety of substrates. This led to the
concept of a cerium oxide self-forming anode based on the principle that
controlled additions of cerium ions to the cryolite bath will form a protective
coating on a suitable anode substrate. The stability of this coating depends
on maintaining a suitable content of cerium in the molten electrolyte and
co-deposing doping elements so as to establish a dynamic equilibrium be-
tween the coating and the dissolved species. With a coating that is electri-
cally conductive and resistant to cryolite bath, many substrates with insuf-
ficient corrosion resistance now qualify for consideration as anode materi-
als in combination with a cerium oxide coating.
It is not yet clear how such a coating forms and resists during electrolysis.
Cathodic deposition of about 2 wt% of cerium in the aluminum dictates that
the aluminum must be refined, and that the recovered cerium is recycled
back into the cryolite melt.
New anode materials, fabricated by micropyretic synthesis, seem to be re-
sistant to cryolite, particularly when protected by the self-forming cerium
oxide coating. Large samples are being prepared for tong-term test instal-

1-17
lation in pilot aluminum cells. Excellent results in 10 kA cells have been

reported.
1.6. Wettable Cathodes
Much money has been invested to develop aluminum-wettable or inert

cathodes (Lit. 1-14 and 1-15) to avoid the following problems during pot
operation.
Magnetic Effects:
Aluminum electrolysis must use a layer of 10-20 cm of aluminum (metal
pad) as cathode for the electrolytic process. Since the carbon bottom
blocks of the electrolysis cell are not wetted by aluminum, the electrolyti-
cally produced aluminum is dispersed into the electrolyte if these carbon
blocks are used as cathode. The liquid aluminum metal is an electric con-
ductor in the magnetic fields of the busbar system. Electromagnetic forces
in the metal pad lead to the magnetic effects namely a deformation of the
electrolyte-metal pad interface (metal upheaving), to a streaming motion
and wavelike oscillation of the metal pad. Because of these magnetic ef-
fects pot operation cannot reduce the anode to cathode distance below
4 cm. A solid cathode material which is wetted by aluminum avoids all
these problems especially since the liquid metal could be drained away
immediately after production.
Chemical Reaction of the Carbon Lining with Bath Constituents:
Liquid electrolyte and aluminum metal penetrate into the carbon lining.
Also the intercalation of sodium with carbon causes swelling and deforma-
tion of the carbon lining. The soft carbon material is attacked by mechani-
cal erosion of the moving bath and by chemical corrosion of the electro-
lyte. Replacing or protecting this carbon material with especially refractory
carbon material would lead to considerable benefits.
Requirements of a Wettable Cathode Material
A material suitable as an inert and wettable cathode lining should satisfy

the following requirements:
• sufficient wettability by molten aluminum,

• satisfactory mechanical strength and resistance to cracking due to thermal or chemi-
cal forces,
• satisfactory adhesion to the underlying carbon bottom,
• electronic conductivity,
• low solubility and low reactivity with aluminum and with cryolite/alumina melts,
• good resistance to penetration by molten electrolyte,
• good resistance to oxidation and to corrosion by any active gases to which it may be
exposed, particularly at high temperatures, on the outside of the cell,
• low thermal conductivity.
To learn about the borides, nitrides and carbides of refractory hard metals
(titanium, zirconium, hafnium, etc.) which have been proposed the inter-

1-18
ested reader should consult the review papers Lit. 1-14 and 1-15. Tita-
nium Boride (TiB2) seems to be the most promising candidate material
because of its price and its properties.
The following main types of arrangements can be found in the literature:
Drained Cathode
Figure 1-15 shows the principle:
Figure 1-15: Schema of a Drained Cathode. The bottom lining is covered with a glued
layer of wettable material. The produced aluminum flows into the central
channel since the cathode is slightly inclined.
The bottom lining is inclined and covered with a wettable material. Several
methods were proposed to apply the material to the lining: glueing of tiles,
chemical vapor deposition, electrodeposition, plasma spraying, glazing or
by supersaturation techniques.
Cathode Elements
Various patents and publications have proposed elements made of wet-
table material which are introduced into conventional cells and are used as
cathode.

1-19
Figure 1-16: Floating Cathode Elements.
Figure 1-16 shows an example of floating cathode elements. Sleeve ele-

ments attached to the carbon lining keep the cathode elements in place. If
the anode blocks touch accidentally the brittle cathode elements they are
not broken since they can be pushed away.
Multipolar Cells
It is a fascinating idea to combine an inert anode material with a wettable
cathode material to form a binary electrode. Such electrodes can be ar-
ranged similar to the Alcoa AlCl3-Cell in a multipolar electrolysis cell (Fig-
ure 1-16).
Figure 1-17: Electrolysis Cell with several Bipolar Electrodes. An intermediate material
separates but connects electrically the anode from the cathode to avoid
chemical reactions.
Several problems must be solved: the anode material must not react
chemically with the cathode material, the products of the electrolysis (alu-
minum and oxygen) must be transported away from the electrodes without
reacting with each other, alumina must be fed to the electrolyte and alumi-
num tapped like in a conventional cell. To what extend these projects were
executed is not known.

1-20
1.7. Literature
Lit. 1-1: E. Bosshard, O. Knaisch, W. Schmidt-Hatting, J. M. Blanc,
“EPT18: The New 180-kA-Pot of Alusuisse”,
Light Metals 1983, 595-605.
Lit. 1-2: H-J. Rebmann, J-C. Bessard, W. Merz,

"Vorrichtung zum Dosieren eines feinkörnigen, rieselfähigen Schüttgutes",
DE 30 49 510 (1980).
Lit. 1-3: P. Bonny, J.-L. Gerphagnon, G. Labouré, M. Keinborg, P. Homsi, B. Langon,

“Process and Apparatus for Accurately Controlling the Rate of Introduction
and the Content of Alumina in an Igneous Electrolysis Tank in the Produc-
tion of Aluminium”,
US 4,431,491 (1984).
Lit. 1-4: W. Merz,

“Förderleitungssystem für pulverförmiges oder körniges Gut”,
DE 30 22 656 (1980).
Lit. 1-5: G.P. Bearne, M. R. Dunn, M. J. Roberts, Y. A. Mohammed,

"The CD200 Project - The Development of a 200 kA Cell Design. From Con-
cept to Implementation",
Lit. 1-6: R. P. Pawlek,

“ALUSAF, Hillside Smelter Project”,
Light Metal Age 1996, 54(7,8), 8-21.
Lit. 1-7: V. A. Kryukovski, G. A. Sirasutdinov, J. Klein, G. Peychal-Heiling,

“International Cooperation and High-Performance Reduction in Siberia”,
JOM 1994, 46(2), 23-25.
Lit. 1-8: A. S. Russel,

"Pitfalls and Pleasures in New Aluminum Process Development"
Met.Trans.B 1981, Vol. 12B, 203-215.
Lit. 1-9: M. J. Bruno,

"Overview of Alcoa Direct Reduction Process"
Light Metals 1984, 1571-1590.
Lit. 1-10: R. L. Troup,

"Feasibilty of an Aluminum-Silicon Blast Furnace"
Light Metals 1984, 1591-1612.
Lit. 1-11: D. T. Stevenson,

"Feasibilty of an Aluminum-Silicon Blast-Arc Furnace"
Light Metals 1984, 1613-1631.

"Inert Anodes for the Primary Aluminium Industry: An Update"

1-21
Lit. 1-13: R. D. Peterson, N. E. Richards, A. T. Tabereaux, O. H. Koski, L. G. Morgan, D.

M. Strachan,
"Results of 100 Hour Electrolysis Test of a Cermet Anode: Operational Re-
sults and Industry Perspective",

"Recent Development of Aluminium-Wettable Cathodes for the Primary
Aluminium Industry Part I",
ALUMINIUM 1995, 71, 732-735.

"Recent Development of Aluminium-Wettable Cathodes for the Primary
Aluminium Industry Part II",
ALUMINIUM 1996, 72, 35-41.

1-22
2. Mass Balance
2. Mass Balance
2.1. Summary ...................................................................... 2-3
2.2. Principal Equation....................................................... 2-3
2.3. Theoretical Production and Consumption................ 2-6
2.4. Reoxidation of Aluminum........................................... 2-8
2.5. Electrolytic Production and Consumption................ 2-9
2.6. Ideal Gas .................................................................... 2-14
2.7. Electrolytic Volume Production ............................... 2-16
2.8. Pearson Waddington Equation ................................ 2-19
2.9. Operating Figures ..................................................... 2-21
2.10. Inert Anodes ........................................................... 2-23
2.11. Literature................................................................. 2-26
2.12. Exercises ................................................................ 2-27

2-1
2. Mass Balance

2-2
2. Mass Balance
2. Mass Balance
2.1. Summary
The meaning of the Principal Equation for the aluminum produc-

tion is shown and the use of molar masses (formerly atomic and
molecular weights) and of moles is explained. Equations to cal-
culate the theoretical productions and consumptions for the
participants of the electrolysis reactions are derived.
Pearson and Waddington have assumed that the loss of current
efficiency is due to the reaction of aluminum with the anodic
carbon dioxide. The expressions for theoretical productions and
consumptions are modified according to the reoxidation reac-
tion with Pearson and Waddington's assumption.
The Ideal Gas Equation connects the pressure, the volume and
temperature of an ideal gas. Expressions for the electrolytic
production of carbon dioxide and carbon monoxide in cubic me-
ter are given.
The Pearson-Waddington Equation relates the composition of
the anode gas to current efficiency. This equation is used to de-
termine the current efficiency of laboratory and technical elec-
trolytic cells from the CO/CO2 ratio of the anode gas.
The conventional operating figures are given to assess the
technical efficiency of a pot room.
2.2. Principal Equation

Alumina (aluminum oxide or Al2O3) is decomposed in an electrolytic proc-
ess to produce aluminum. At the positive electrode, the anode, oxygen is
formed which reacts immediately with the carbon anode to carbon dioxide.
Aluminum is deposited at the negative electrode, the cathode, which is the
already produced aluminum pool. This process is described by the
Principal Equation:
2Al2O3 + 3C = 4Al + 3 CO2 (2–1)
The Principal Equation means that two moles Al2O3 react with three
moles C to four moles Al and three moles CO2. The Mole is defined as the
amount of substance (symbol: n, unit: mol) which contains as many ele-
mentary entities as there are atoms in 0.012 kg 12C (Lit. 2-2, p.40). This
number NA=6.022×1023 is called Avogadro Constant (Table 1, p.1-1 of
Lit. 2-3).

2-3
2. Mass Balance
During chemical processes particles (atoms, molecules) are reacting with

each other. By using mole expressions formulating chemical reactions be-
come independent of the properties of the reacting substances and are
valid (with reservations) for every system.
For technical applications a more practical unit namely the molar mass of
a substance is applied. The Molar Mass of a substance is defined as the
mass of the substance divided by its amount of substance:
M =m (2–2)
n
with the meaning of the symbols:
M: molar mass (g/mol),
m: mass of substance (g),
n: number of moles.
The molar mass of a substance B of definite chemical composition or of an

atom X is given by
M B = M r (B) (2–3)
M X = Ar (X)
MB: molar mass of compound B (10-3 kg/mol, kg/mol, g/mol),
Mr(B): relative molecular mass of compound B,
Mx: molar mass of atom X (10-3 kg/mol, kg/mol, g/mol),
Ar(X): relative atomic mass of atom X.
Where the Relative Atomic Mass (formerly atomic weight) is defined as

the ratio of the average mass per atom of an element to 1/12 of the mass
of an atom of the nucleide 12C and the Relative Molecular Mass (formerly
molecular weight) as ratio of the average mass per molecule or specified
entity of a substance to 1/12 of the mass of an atom of the nucleide 12C.
You find relative atomic and molecular masses of elements and com-
pounds which are important for the Hall-Héroult-process in Table 2-1 and
Table 2-2.

2-4
2. Mass Balance
Table 2-1: Relative Atomic Masses of some Elements.
Table 2-2: Relative Molecular Masses of some Chemical Compounds.
Example 2-1: Determination of the Relative Molecular Mass of Alumina.

To determine the relative molecular mass of alumina Mr(Al2O3) you find in
Table 2-1 the relative atomic mass of aluminum Mr(Al)=26.9815 and of
oxygen Mr(O)=15.9994. The relative molecular mass is then:
c h a f bg
M r Al2 O 3 = 2 ⋅ M r Al + 3 ⋅ M r O = 2 × 26.9815 + 3 × 15.9994 = 1019612
.

2-5
2. Mass Balance
2.3. Theoretical Production and Consumption

Faradays’s Laws of electrolysis are saying that
1. the amount of substances deposited or dissolved during an electrolytic

process is proportional to the quantity of electricity passed through the
electrolytic cell.
2. One gram equivalent weight of matter is deposited or dissolved on each
electrode for 96485 Coulomb of electricity charge passed through the
electrolyte.
The Equivalent Weight of a substance is its atomic or molecular weight

divided by its electrochemical valence e.g. the ions Na+, 1/2Ca2+, and
1/3Al3+ are electrochemically equivalent. The Gram Equivalent Weight is
the mass in grams numerically equal to the equivalent weight.
The unit of the electric charge is the Coulomb (1 C = 1 A·s). One mole of
electrons has an electric charge of 96485 C. This value F = 96485 C/mol
is called the Faraday Constant (Table 1 p.1-1 of Lit. 2-3).
During electrolysis the electric charge of the aluminum ion is changed from
positive three plus to electrical neutral. Accordingly three electrons are
used per aluminum ion or three moles of electrons per one mole of alumi-
num:
Al3+ + 3 e- = Al (2–4)
One writes for the
Theoretical Consumptions (TC)

of aluminum oxide:
TC Al O
= 0.6341⋅I ⋅t (2–5)
2 3
of carbon:
TC C = 01120
. ⋅I ⋅t (2–6)
and
Theoretical Productions (TP)

of aluminum:
TPAl = 0.3356 ⋅I ⋅t (2–7)

2-6
2. Mass Balance
of carbon dioxide:
TPCO = 0.4105 ⋅I ⋅t (2–8)

2
TC: theoretical consumption (kg),
TP: theoretical production (kg),
I: electric current (kA),
t: time (h).
Equation 2-7 for TPAl is derived in the following way (the derivations of the
other equations you find in ElyseSem):
To discharge four aluminum ions (4Al3+) according to the Principal Equation
2-1 4×3×F of electric charge is necessary. When a current I during the time t
is flowing through the electrolytic cell the theoretical production of aluminum
is:
4
TPAl = ⋅ M Al ⋅ I ⋅ t
4 ⋅ 3 ⋅F
where MAl is the molar mass of aluminum. Putting in the values
4 26.9815
TPAl = × 3
× I × 10 3 × t × 3600
12 × 96485 10
one gets TPAl (Equation 2–7) in kg when I is in kA and t in h :
TPAl = 0.3356 ⋅ I ⋅ t
According to the Principal Equation 2-1 the following relations must hold:
TC Al O + TC C = TPAl + TPCO
2 3 2 (2–9)
0.6341 + 0.1120 = 0.3356 + 0.4105
Example 2-2: Theoretical Productions and Consumptions

An electrolytic pot is operated at 300 kA. How much aluminum and carbon
dioxide is it producing and how much alumina and carbon is it consuming
theoretically in 24 hours?
You write for the theoretical consumption of aluminum oxide (Equation

2-5):
TC Al O = 0.6341⋅ I ⋅ t = 0.6341× 300 × 24 = 4566 kg
2 3
for the theoretical consumption of carbon (Equation 2-6):

TC C = 0.1120 ⋅ I ⋅ t = 0.1120 × 300 × 24 = 806 kg
for the theoretical production of aluminum (Equation 2-7):

TPAl = 0.3356 ⋅ I ⋅ t = 0.3356 × 300 × 24 = 2416 kg
and for the theoretical production of carbon dioxide (Equation 2-8):

2-7
2. Mass Balance
TPCO = 0.4105 ⋅ I ⋅ t = 0.4105 × 300 × 24 = 2956 kg .

2
Please see Example 2-3 especially concerning the use of ElysePrg.
2.4. Reoxidation of Aluminum

In practice an electrolytic pot produces always less aluminum than is cal-
culated according to Faraday's Laws. The current efficiency is the ratio of
actually produced aluminum over theoretically produced aluminum. This
value is called
Fractional Current Efficiency:

PAl PAl
η= = (2–10)
TPAl 0.3356 ⋅I ⋅t
η: fractional current efficiency,
PAl: actually produced aluminum (kg),
TPAl: theoretically produced aluminum (kg),
t: Time (h).
Normally the current efficiency is expressed in percent the
Current Efficiency in Percent:

CE=100·η (2–11)
CE: current efficiency in percent.
The electrolyte dissolves aluminum which is transported to the anode.

There it reacts with carbon dioxide to alumina according to the
Reoxidation Reaction Equation:

2Al + 3CO2 = Al2O3 + 3CO (2–12)
Pearson and Waddington (Lit. 2-4) assume that the loss in current effi-
ciency is only due to the reaction of aluminum with CO2.
To the Principal Equation
2Al2O3 + 3C = 4Al + 3 CO2

the Reoxidation Reaction Equation is added
2·(1-η)[2Al + 3CO2 = Al2O3 + 3CO]
considering that (1-η) fractions of aluminum are lost if η fractions are pro-
duced which results in the Electrolysis Equation:

2-8
2. Mass Balance
2ηAl2O3 + 3C = 4ηAl + 3(2η-1)CO2 + 6(1-η)CO (2–13)
Remark: You can add (or subtract) chemical equations like mathematical
equations. You have to consider, however, that the sign has to be
changed when you shift one component to the other side of the equation:
2.5. Electrolytic Production and Consumption

According to the Electrolysis Equation 2-13 e.g. to the theory of Pearson
and Waddington the expressions for the theoretical productions and con-
sumptions (Equation 2-5 to 2-8) must be modified. One writes for the
Electrolytic Consumptions (EC)

of aluminum oxide:
EC Al O
= 0.6341⋅I ⋅t ⋅η (2–14)
2 3
of carbon:
EC C = 01120
. ⋅I ⋅t (2–15)
and
Electrolytic Productions (EP)

of aluminum:
EPAl = 0.3356 ⋅I ⋅t ⋅η (2–16)

of carbon dioxide:
EPCO = 0.4105 ⋅I ⋅t ⋅(2η − 1) (2–17)

2
of carbon monoxide:
EPCO = 0.5226 ⋅I ⋅t ⋅(1− η) (2–18)

EC: electrolytic consumption (kg),
EP: electrolytic production (kg),
t: time (h),
η: fractional current efficiency.
You derive the Equation 2-16 for EPAl form Equation 2-7:

2-9
2. Mass Balance
4
EPAl = TPAl ⋅ η = ⋅ M Al ⋅ I ⋅ t ⋅ η
4 ⋅ 3 ⋅F
4 26.9815
EPAl = × 3
× I × 10 3 × t × 3600 × η = 0.3356 ⋅ I ⋅ t ⋅ η
12 × 96485 10
and for EPCO (2-18) in the following way (the derivations of the other equa-
tions you find in ElyseSem):
To produce 6(1-η) carbon monoxide molecules according to the Electrolysis

Equation 4×3×F of electric charge is necessary. When a current I during the
time t is flowing through the electrolytic cell the production of carbon dioxide
is:
6
EPCO = ⋅ M CO ⋅ I ⋅ t ⋅ (1− η)
4 ⋅ 3 ⋅F
where MCO is the molar mass of carbon monoxide (Table 2-2).

Putting in the values
6 28.0105
EPCO = × 3
× I × 10 3 × t × 3600 × (1− η)
12 × 96485 10
one gets EPCO in kg when I is in kA and t in h and η is the fractional current

efficiency:
EPCO = 0.5226 ⋅ I ⋅ t ⋅ (1− η)
The calculated values of EPCO2 and EPCO are in kilograms. If you want
them in cubic meters please use Equation 2-30 and 2-31.
The practice uses also specific values e.g. values related to a reactant of
the Principal Equation, normally to one kilogram of aluminum. One can
write for the
Specific Electrolytic Consumptions (SEC)

of aluminum oxide:
SEC Al O
= 18895
. (2–19)
2 3
of carbon:
SEC C = 0.3338 (2–20)
η
and Specific Electrolytic Productions (SEP)
of carbon dioxide:
2η − 1
SEPCO = 12233
. ⋅ (2–21)
2 η
of carbon monoxide:

2-10
2. Mass Balance
1− η
SEPCO = 15572
. ⋅ (2–22)
η
where the SEC and SEP are in fractions (kg/kg) and η is the fractional cur-
rent efficiency. The values of SEPCO2 and SEPCO are calculated by dividing
the kilograms of CO2 and CO by the kilograms aluminum. If you want them
in cubic meters per kilogram please use Equation 2-32 and 2-33.
Remark: It is the habit to write the units of specific values for instance for
SECC like kgC/kgAl. According to (Lit. 1-1 and Lit. 2-2) it is not acceptable
to mix any information concerning the quantity with the unit.
To derive SECAl2O3 (Equation 2-19) you write:
EC Al O 2 4 ⋅ 3 ⋅F 1
SEC Al O = 2 3
= ⋅ M Al O ⋅ I ⋅ t ⋅ η ⋅ ⋅
2 3 EPAl 4 ⋅ 3 ⋅F 2 3 4 M Al ⋅ I ⋅ t ⋅ η
1 M Al2O3 1 1019612
.
SEC Al O = ⋅ = × = 18895
.
2 3 2 M Al 2 26.9815
and for SECCO (Equation 2-21) you write (the derivations of the other
equations you find in ElyseSem):
EPCO 6 4 ⋅ 3 ⋅F 1
SEPCO = = ⋅ M CO ⋅ I ⋅ t ⋅ (1− η) ⋅ ⋅
EPAl 4 ⋅ 3 ⋅F 4 M Al ⋅ I ⋅ t ⋅ η
3 M CO 1− η 3 28.0105 1− η 1− η
SEPCO = ⋅ ⋅ = × × = 15572
. ⋅
2 M Al η 2 26.9815 η η
According to the Electrolysis Equation 2-13 the following relations must
hold (see Equation 2-9):
ECAl2O3 + ECC = EPAl + EPCO2 + EPCO

(2–23)
0.6341×η + 0.1220 = 0.3356×η + 0.4105×(2η-1) + 0.5226×(η-1)
and
SECAl2O3 + SECC = 1+ SEPCO2 + SEPCO

(2–24)
1.8895×η + 0.3338 = η + 1.2233×(2η-1) + 1.5572×(1-η)
The following examples show how to apply the expressions derived in this
chapter. Example 2-3 calculates correspondingly to Example 2-2 the elec-
trolytic productions and consumption values.
Example 2-4 investigates an experiment with a laboratory cell and deter-
mines the current load and current efficiency of a laboratory cell.

2-11
2. Mass Balance
Example 2-3: Electrolytic Production and Consumption

Corresponding to the Example 2-2 an electrolytic pot is running at 300 kA
with a current efficiency in percent (Equation 2-11) CE = 95 %. How much
aluminum, carbon dioxide and carbon monoxide is it producing and how
much alumina and carbon is it consuming in 24 hours? Calculate also the
specific values.
You write for the electrolytic consumption (Equation 2-14) and specific
electrolytic consumption (Equation 2-19) of aluminum oxide :
EC Al2O 3 = 0.6341⋅ I ⋅ t ⋅ η = 0.6341× 300 × 24 × 0.95 = 4338 kg
SEC Al2O 3 = 1890
. kg / kg
for the electrolytic consumption (Equation 2-15) and specific electrolytic
consumption (Equation 2-20) of carbon :
EC C = 0.1120 ⋅ I ⋅ t = 0.1120 × 300 × 24 = 806 kg
0.3338 0.3338
SEC C = = = 0.351 kg / kg
η 0.95
for the electrolytic production (Equation 2-16) of aluminum :

EPAl = 0.3356 ⋅ I ⋅ t ⋅ η = 0.3356 × 300 × 24 × 0.95 = 2295 kg
for the electrolytic production (Equation 2-17) and specific electrolytic pro-
duction (Equation 2-21) of carbon dioxide:
EPCO 2 = 0.4105 ⋅ I ⋅ t ⋅ (2η − 1) = 0.4105 × 300 × 24 × (2 × 0.95 − 1) = 2660 kg
2η − 1 2 × 0.95 − 1
SEPCO 2 = 12233
. ⋅ = 12233
. × = 1159
. kg / kg
η 0.95
and for the electrolytic production (Equation 2-1) and specific electrolytic
production (Equation 2-22) of carbon monoxide:
EPCO = 0.5226 ⋅ I ⋅ t ⋅ (1− η) = 0.5226 × 300 × 24 × (1− 0.95) = 188 kg
1− η 1− 0.95
SEPCO = 15572
. ⋅ = 15572
. × = 0.082 kg / kg
η 0.95
If you want the gas production in cubic meters see Example 2-7.
Click in the List Examples and Exercises Window of ElysePrg

(Chap. 18.14) on Exp. 2-3 and ElysePrg executes the calculation of
Exp. 2-3 and shows the following calculation page:

2-12
2. Mass Balance
For more information about ElysePrg please see Chap. 18 and specifically
about these calculations see Chap. 18.3.1.
Example 2-4: Current Load and Current Efficiency of a Laboratory Cell

During 8 hours an experiment was conducted in a laboratory cell. 45.9 g of
aluminum were produced and 22.4 g of carbon electrolytically consumed.
What was the current load and the current efficiency?
Transforming the relation 2-20 for the specific electrolytic consumption of

carbon one calculates the current efficiency in percent (Equation 2-11):
EPAl 45.9
CE = 100 × 0.3338 × = 33.39 × = 68.4 %
EC C 22.4
Using Equation 2-16 for the elctrolytic production of aluminum one finds for
the current load:
EPAl 45.9
I= = = 25 A
0.3356 ⋅ η ⋅ t 0.3356 × 0.684 × 8

2-13
2. Mass Balance
2.6. Ideal Gas

An ideal gas is regarded as one for which the molecular attraction is negli-
gibly small and in which the actual volume of the molecules is small in
comparison with the space they inhibit. The Ideal Gas Equation describes
the state of such an ideal gas:
p·V = n·R·T (2–25)

p: Pressure (Pa, atm, mmHg),
V: Volume (m3, cm3, l, ft3),
n: Number of moles (mol),
R: Gas Constant (J/(mol·K)),
T: Temperature (K).
The gas pressure and temperature are measured in different units. The
following tables inform how to convert those units:
Table 2-3: Conversion of Temperatures (Lit 2-3, p. 1-33).
Table 2-4: Conversion Factors for Pressure (Lit 2-3, p.1-35).

2-14
2. Mass Balance
The value of the molar Gas Constant (Lit 2-3, p.1-43) is:
F Pa ⋅m3 J I
R = 8.314510 GG or J (2–26)
H K ⋅mol K ⋅molJK
Table 2-5 shows the values of the gas constant in different unit systems:
Table 2-5: The table Values of the Gas Constant
The volume of one mole of ideal gas at the Standard State namely a tem-
perature of T0 = 273.15 K (0 °C) and a pressure of p0 = 101 325 Pa
(1 atm) is called Molar Volume of the ideal gas . Its value is:
R ⋅T 0 8.31451× 273.15 m3
Vm = = = 0.002 2414
p0 101325
. × 10 5 mol
3
Vm = 0.022414 m = 22.414 L (2–27)
mol mol
If you want to calculate the volume of an ideal gas at a different tempera-

ture (t) or pressure (p) other than an initial state (t0, p0) use the
Conversion Equation for the Volume of an Ideal Gas:
V t,p =V t ⋅ . ⋅ p0
t + 27315 (2–28)
,p
0 0 t + 27315
0
. p
Vt 0 ,p0 : Volume of the ideal gas at the initial (t0, p0),
state
Vt,p: Volume of the ideal gas at the final state (t,p),
t0,t:: Initial, final temperature (°C),
p0, p: Initial, final pressure.
The following examples show the application of these equations:

Example 2-5: Gas Volume at Standard Conditions
An electrolytic pot is producing 1850 cubic meters of CO2 at 95 °C. How
much gas is the pot producing at standard conditions?
Using Equation 2-28 for the volume of an ideal gas and the values for the
standard state of an ideal gas one calculates:

2-15
2. Mass Balance
e j e
V t = 0 o C = V t = 95 o C ⋅ j 95273.15
+ 273.15
= 1850 ×
273
368
= 1372 m3
Example 2-6: Gas Volume after Change of Temperature and Pressure

Surfing the WW (World Wide Web: http:/www-hpcc.astro.washington.edu
/mirrors/suremath/gasLaw.html) you find the following problem:
If 333 ml of an ideal gas at a temperature of 25 °C and a pressure of 750
torr has its temperature lowered to -11 °C and its pressure to 730 mmHg,
what will its new volume be in milliliters?
Using the Equation 2-28 and finding in Table 2-4 that torrs are equal to mil-
limeters of mercury (mmHg) one calculates:
t + 273.15 p 0 −11+ 273 750

V t,p = V t 0 ,p0 ⋅ ⋅ = 333 × × = 301ml
t 0 + 273.15 p 25 + 273 730
2.7. Electrolytic Volume Production

Using Equation 2-2 one writes for the number of moles of carbon dioxide
produced during electrolysis
m CO EPCO (2–29)
n CO = 2 = 2
2 M CO M CO
2 2
nCO2: number of moles of carbon dioxide,
mCO2: mass of carbon dioxide,
MCO2: molecular mass of carbon dioxide,
EPCO2: electrolytic production of carbon dioxide.
Using Equation 2-17 and 2-18 (electrolytic productions EP) and the Equa-
tion 2-25 (ideal gas) one finds the following relations for the
Electrolytic Volume Productions (VP)

of carbon dioxide:
VPCO = 0.2091⋅I ⋅t ⋅ (2η − 1) (2–30)

2
and of carbon monoxide:
VPCO = 0.4181⋅I ⋅t ⋅ (1− η) (2–31)

VPCO2: electrolytic volume production of carbon dioxide (m3),
VPCO: electrolytic volume production of carbon monoxide (m3),
t: Time (h),
at standard conditions: T0 = 273.15 K (0 °C),
p0 = 101 325 Pa (1 atm).

2-16
2. Mass Balance
Derivation of Equation 2-30:
R ⋅T EPCO2 R ⋅T 3 R ⋅T
VPCO = n CO ⋅ = ⋅ = ⋅ ⋅ I ⋅ t ⋅ (2η − 1)
2 2 p M CO p 4 ⋅ 3 ⋅F p
2
3
3600 × 10 8.31451× 273.15
VPCO = × × I × t × (2η − 1) = 0.2091⋅ I ⋅ t ⋅ (2η − 1)
2 4 × 96485 101325
. × 10 5
and of Equation 2-31:
R ⋅T EPCO R ⋅T 6 R ⋅T
VPCO = n CO ⋅ = ⋅ = ⋅ ⋅ I ⋅ t ⋅ (1− η)
p M CO p 4 ⋅ 3 ⋅F p
3600 × 10 3 8.31451× 273.15
VPCO = × × I × t × (1− η) = 0.4182 ⋅ I ⋅ t ⋅ (1− η )
2 × 96485 101325
. × 10 5
The production values of VPCO2 and VPCO are in cubic meters (m3). If you
want them in kilograms (EPCO2, EPCO) please use Equation 2-21 and 2-22
of the electrolytic productions.
Similarly to the specific electrolytic consumptions and productions one

writes using Equation 2-21 and 2-22 for the
Specific Electrolytic Volume Productions (SVP):

of carbon dioxide:
2η − 1
SVPCO = 0.6230 ⋅ (2–32)
2 η
and of carbon monoxide:
1− η
SVPCO = 12460
. ⋅ (2–33)
η
SVPCO2: specific electrolytic volume production of carbon dioxide (m3/kg),
SVPCO: specific electrolytic volume production of carbon monox- (m3/kg),
ide
at standard conditions: T0 = 273.15 K (0 °C),
p0 = 101 325 Pa (1 atm).
Derivation of Equation 2-32:
1 R ⋅T 1 R ⋅T 3 1 R ⋅T 2η − 1
SVPCO = ⋅ nCO ⋅ = SEPCO . ⋅ = ⋅ ⋅ ⋅
2 EPAl 2 p 2 M CO p 4 M Al p
2
η
3
3 10 8.31451× 273.15 2η − 1 2η − 1
SVPCO = × × × = 0.6230 ⋅
2 4 26 ⋅ 9815 101325
. × 10 5 η η
and of Equation 2-32:

2-17
2. Mass Balance
1 R ⋅T 1 R ⋅T 3 1 R ⋅T 1− η
SVPCO = ⋅ n CO ⋅ = SEPCO . ⋅ = ⋅ ⋅ ⋅
EPAl p M CO p 2 M Al p η
3 10 3 8.31451× 273.15 1− η 1− η
SVPCO = × ⋅ × = 12460
. ⋅
2 26 × 9815 101325
. × 10 5 η η
The calculated values of SVPCO2 and SVPCO are in cubic meters per kilo-
gram. If you want them in kilograms per kilogram please use Equa-
tion 2-21 and 2-22.
Example 2-7: Electrolytic Production of CO2 and CO in Cubic Meters
Corresponding to the Example 2-3 an electrolytic pot is operated at 300
kA with a current efficiency (Equation 2-11) of CE = 95 %. How much car-
bon dioxide and carbon monoxide is it producing at standard conditions in
24 hours? Calculate also the specific values.
You write for the electrolytic volume production (Equation 2-30) and spe-
cific electrolytic volume production (Equation 2-32) of carbon dioxide:
VPCO 2 = 0.2091⋅ I ⋅ t ⋅ (2η − 1) = 0.2091× 300 × 24 × (2 × 0.95 − 1) = 1355 m 3
2η − 1 2 × 0.95 − 1
SVPCO 2 = 0.6230 ⋅ = 0.6230 × = 0.590 m 3 / kg
η 0.95
and for the specific electrolytic volume production (Equation 2-31) and
specific electrolytic volume production (Equation 2-33 ) of carbon monox-
ide:
VPCO = 0.4181⋅ I ⋅ t ⋅ (1− η ) = 0.4181× 300 × 24 × (1− 0.95) = 151 m 3
1− η 1− 0.95
SVPCO = 12460
. ⋅ = 1246
. × = 0.066 m 3 / kg
η 0.95
If you want the gas production in kilograms see the Example 2-3.
Click in the List Examples and Exercises Window of ElysePrg (Chap.
18.14) on Exp. 2-7 and ElysePrg executes the calculation and shows the
following calculation page:

2-18
2. Mass Balance
about these calculations see Chap. 18.4.1 and Chap. 18.4.2.
2.8. Pearson Waddington Equation

The content of the components in a system (A, B) are given in fractions.
The following values are used:
Mass Fraction
mA mB
wA = wB = (2–34)
m A + mB m A + mB
wA, wB: mass fractions of component A, B,
mA, mB: masses of component A, B.
Mole Fraction
nA nB
xA = xB = (2–35)
n A + nB n A + nB
xA, xB: mole fractions of component A, B,
nA, nB: number of moles of component A, B.
Volume Fraction
VA VB (2–36)
ϕA = ϕB =
V A +VB V A +VB
ϕA, ϕB: volume fractions of component A, B,
VA, VB: volume of component A, B.
For the volume fraction and for the corresponding mole fractions of carbon
dioxide in the anode gas one writes by using the Ideal Gas Equation 2-25:
V CO n CO (2–37)
ϕ CO = 2 = 2 = x CO
2 V CO +V CO n CO + n CO 2
2 2
ϕCO2: volume fraction of carbon dioxide in the anode gas,
VCO2: volume of carbon dioxide in the anode gas,
volume of carbon monoxide in the anode gas,
nCO2: number of moles of carbon dioxide in the anode gas,
nCO: number of moles of carbon monoxide in the anode gas,
xCO2: mole fraction of carbon dioxide in the anode gas.
Taking the values from the Electrolysis Equation 2-13 one continues:
2η − 1
ϕ CO = x CO = (2–38)
2 2 FH 2η −1IK + 2 ⋅ FH1− ηIK
2-19
2. Mass Balance

xCO2: mole fraction of carbon dioxide in the anode gas,
Rearranging for the current efficiency η yields the
Pearson Waddington Equation (Lit. 2-4):
ϕ CO 100 ⋅ϕ CO (2–39)
η = 0.5 + 2 or CE = 50 + 2
2 2
CE: current efficiency in percent.
or
ϕ CO 100 ⋅ϕ CO (2–40)
η = 1− or CE = 100 −
2 2
ϕCO: volume fraction of carbon monoxide in the anode gas,
CE: current efficiency in percent
by using
ϕ CO = 1− ϕ CO (2–41)
2
Remarks: According to the Pearson Waddington Equation the current effi-

ciency can not be less than 50%. The content of carbon monoxide in the
anode gas is about half the loss of current efficiency.
The Pearson Waddington Equation 2-39 is used to determine the current

efficiency from the composition of the pot gases. Especially for laboratory
cells this method was applied successfully. Chapter 9.4 (p.336-344) of
Lit. 2-5 gives a review about the current efficiency in laboratory cells, Lit.
2-6 and Lit. 2-7 are more recent examples.
Also the current efficiency of industrial electrolytic pots (Lit. 2-5, Chapter
9.5 p. 344-351) was examined frequently by analyzing samples of the an-
ode gas taken through holes in the anodes ( Lit. 2-8 is a recent example).
Also samples of the exhaust gases ( Lit. 2-9) were used for current effi-
ciency determinations.

2-20
2. Mass Balance
2.9. Operating Figures

The operating figures normally contain the consumptions and productions
as specific values e.g. in relation to one kilogram of produced aluminum.
The consumption or production value of an item over a time period is di-
vided by the amount of aluminum produced in the same time period.
Several values of Anode Consumption (kg/kg or kg/ t) are defined:
Anode Gross Consumption (kg/kg): AGC,

Anode Net Consumption (kg/kg): ANC,
Anode Butts (kg/kg): AB,
Anode Excess Consumption (kg/kg): AEC.
The following relations are holding:
AGC = ANC + AB (2–42)

and
ANC = AEC + SECC (2–43)

with SECC: electrolytic specific consumption of carbon (Equation 2-20).
Figure 2-1 shows graphically the relation between the anode consump-
tions.
Figure 2-1: Graphical representation of the relation between the different types of anode
consumptions. AGC is the anode gross consumption, AB are the anode butts,
ANC is the anode net consumption, AEC the anode excess consumption and
SECC the electrolytic specific consumption (Equation 2-20).

2-21
2. Mass Balance
Table 2-6 lists the specific consumptions which are normally determine for
technical pots:
Table 2-6: Technical Specific Consumptions
Remark: It is the habit to write the units of specific values for instance
kWh/kgAl. According to (Lit.2-1 and Lit. 2-2) it is not acceptable to mix any
information concerning the quantity with the unit.
Table 2-7 shows an example of monthly operating figures:

Table 2-7: Example of monthly operating figures.

2-22
2. Mass Balance
2.10. Inert Anodes

Operating an electrolytic pot with inert anodes means that the anodic oxy-
gen does not react with the anode material (see Chap 1.5). Corresponding
to the electrolysis equation (Equ. 2-13) of the conventional Hall-Héroult
process you write for an electrolysis process that uses inert anodes the
Electrolysis Equation with:
2ηAl2O3 = 4ηAl + 3ηO2 (2-44)

The relations for the electrolytic consumption of alumina (Equ. 2-14) and
the electrolytic production of aluminum (Equ. 2-16) are the same as for the
conventional electrolysis. However the Electrolytic Production of Oxy-
gen is different:
EPO = 0.2985 ⋅ I ⋅ t ⋅η (2-45)

2
with the derivation:
1
EPO2 = ⋅ MO2 ⋅ I ⋅ t ⋅ 3 ⋅ η
4 ⋅ 3 ⋅F
1 31.9988 (2-46)
EPO2 = × × I × 103 × t × 3600 × 3 × η
12 × 96485 10 3
EPO2 = 0.2985 × I × t × η
as is the Specific Electrolytic Production of Oxygen:
SEPO = 0.8895 (2-47)

2

2-23
2. Mass Balance
EPO2 1 4 ⋅ 3 ⋅F 1
SEPO2 = = ⋅ MO2 ⋅ I ⋅ t ⋅ 3 ⋅ η ⋅ ⋅
EPAl 4 ⋅ 3 ⋅F 1 MAl ⋅ I ⋅ t ⋅ 4 ⋅ η
(2-48)
3 MO 3 31.9988
SEPO2 = ⋅ 2 = × = 0.8895
4 MAl 4 26.9815
Similar to Equ. 2-30 and 2-31 you write for the Volume Production of
Oxygen
VPO = 0.2091⋅ I ⋅ t ⋅η (2-49)

2
R ⋅ T EPO2 R ⋅ T 1 R ⋅T
VPO2 = nO2 ⋅ = ⋅ = ⋅ ⋅ I ⋅ t ⋅ 3 ⋅η
p MO2 p 4 ⋅ 3 ⋅F p
(2-50)
1 8.31451× 273.15
VPO2 = × × I × 103 × t × 3600 × 3 × η
12 × 96485 1.01325 × 10 5
VPO2 = 0.2091× I × t × η
and corresponding to Equ. 2-32 and 2-33 you write for the Specific Vol-
ume Production of Oxygen

2-24
2. Mass Balance
SVPO = 0.6230 (2-51)

2
1 R ⋅T 1 R ⋅T 3 1 R ⋅T
SVPO2 = ⋅ nO2 ⋅ = SEPO2 ⋅ ⋅ = ⋅ ⋅
EPAl p MO2 p 4 MAl p
(2-52)
3 103 8.31451× 273,15
SEPO2 = × × = 0.6230
4 26.9815 1.01325 × 105

2-25
2. Mass Balance
2.11. Literature
Lit. 2-1: International Organisation for Standardization,
"ISO STANDARDS HANDBOOK: Quantities and Units",
ISBN 92-97-10185-4, Third Edition 1984.
Lit. 2-2: B. N. Taylor

"Guide for the Use of the International System of Units (SI)",
NIST Special Publication 811,
Edition: National Institute of Standards and Technology,
United States Department of Commerce,Technology Administration.
The document may be downloaded from:
http://physics.nist.gov/Pubs/SP811/contents.full.html
Lit. 2-3: D. R. Lide (Editor-in-Chief)

"Chemical Rubber Company Handbook of Chemistry and Physics",
77th Edition (1996-97).
CRC Press, Boca Raton, New York, London, Tokyo;
Lit. 2-4: T.G. Pearson, J. Waddington

"Electrode Reaction in the Aluminium Reduction Cell",
Discuss.Faraday Soc. (1947), 1, 307.
Lit. 2-5: K. Grjotheim, C. Krohn, M. Malinovský, K. Matiašovský, J. Thonstad

"Aluminium Electrolysis, Fundamentals of the Hall-Héroult-Process",
2nd Edition (1982).
Aluminium Verlag, Düsseldorf;
Lit. 2-6: M. M. R. Dorreen, M. M. Hyland, B. J. Welch

"An Improved Method for Current Efficiency Determination in a Laboratory
Aluminium Cell",
Light Metals 1996, 235.
Lit. 2-7: A. Silný, T. A. Utigard

"Determination of the Factors which Control the CO/CO2 Ratio of the An-
ode Gas",
Lit. 2-8: K. A. Paulsen, J. Thonstad, S. Rolseth, T. Ringstad

"Current Efficiency as a Function of the Contents of Alumina and CaF2 in
Industrial Electrolysis Cells",
Lit. 2-9: M. J. Leroy, T. Pelekis, J.-M. Jolas

"Continuous Measurement of Current Efficiency, by Mass Spectrometry,
on a 280 kA Prototype Cell",

2-26
2. Mass Balance
2.12. Exercises
2-1: Radius of the Aluminum Atom.
2-2 : Fluorine Content of Bath Constituents.
2-3: Gas Production.
2-4: Alcoa Smelting Process.
2-5: Anode Consumption in 1 MAh.
2-6: Sodium Oxide in Alumina.
2-7: Reaction of Sodium Carbonate with Aluminum Fluoride.

2-27
2. Mass Balance
Exercise 2-1:
Estimation of Atomic Radius of Aluminum
Estimate the radius of the aluminum atom by using the Avogadro Constant
and the density of aluminum ρ = 2.4 g /cm3.
In 27 grams of aluminum there are 6.022·1023 atoms.

Using
m
ρ=
V
for the density you can write for the volume of the cube which contains
one aluminum atom
m 27 −23 3
V = = = 1.8 ⋅ 10 cm
ρ 2.4 ⋅ 6 ⋅ 10
23
and for the lateral length of this cube
. ⋅ 10 −23 = 2.6 ⋅ 10 −8 cm .
l = 3 18
This fits quite well with an atomic radius of about r ≈ 1.3 Å (Ångstrom:
1 Å = 10-10 m) found in the literature.

2-28
2. Mass Balance
Exercise 2-2:
Fluorine Content of Bath Constituents
Determine the fluorine content of the constituents of the electrolyte: so-

dium cryolite, aluminum fluoride, sodium fluoride and calcium fluoride.
Using Table 2-1 for the relative atomic masses (formerly atomic weights)
and Table 2-2 for the relative molecular masses (formerly molecular
weights) you write:
for Sodium Cryolite
6 × Ar ( F ) 6 × 18.9984
cF = × 100 = × 100 = 54.3 %
M r ( Na 3 AlF6 ) 209.9413
for Aluminum Fluoride
3 × Ar ( F ) 3 × 18.9984
cF = × 100 = × 100 = 67.9 %
M r ( AlF3 ) 83.9768
for Sodium Fluoride
1 × Ar ( F ) 18.9984
cF = × 100 = × 100 = 45.2 %
M r ( NaF ) 41.9882
and for Calcium Fluoride
2 × Ar ( F ) 2 × 18.9984
cF = × 100 = × 100 = 48.7 %
M r ( CaF2 ) 78.0752

2-29
2. Mass Balance
Exercise 2-3:
Gas Production
An electrolytic pot is running with a line current of I = 116.5 kA (CE =

95.2 %). How much CO/CO2 is the pot producing in one hour (temperature
of the gas 200 °C, normal pressure)? Verify the Pearson-Waddington
Equation.
You write using Equation 2-30 and 2-31 for the volume production of car-
bon monoxide/dioxide and Equation 2-28 for volume conversion:
200 + 273.15 3
VPCO = 0.2091 × 116.5 × ( 2 × 0.952 − 1) × = 38 m
2
273.15
200 + 273.15 3
VPCO = 0.4181 × 116.5 × (1 − 0.952 ) × = 4m
273.15
To verify the Pearson-Waddington Equation 2-39 you write:
38
ϕ CO = = 0.905
2
38 + 4
90.5
CE = 50 + = 95.2 %
2
18.14) on Exs. 2-3 and ElysePrg executes the calculation of Exs. 2-3 and
shows the following calculation page:
about these calculations see Chap. 18.4.1 and Chap. 18.4.2.

2-30
2. Mass Balance
Exercise 2-4:
Alcoa Smelting Process
The Alcoa Smelting Process produced aluminum by electrolytically de-

composing aluminum chloride. Determine the current intensity of the elec-
trolytic cell. The cell produced 15000 t/y aluminum with a current efficiency
of CE = 90 %. The number of bipolar electrodes you find in Figure 1-12 .
The electrolytic production EPAl of aluminum per day was:
15000
EPAl = = 411
. t
365
Rewriting Equation 2-16 you calculate the current intensity I = 473 kA:
1 EPAl 41100
I = ⋅ = = 472.5 kA
12 0.3356 ⋅ η ⋅ t 12 × 0.3356 × 0.90 × 24

18.14) on Exs. 2-4 and ElysePrg executes the calculation of Exs. 2-4 and
shows the following calculation page:
Obviously you may only use the aluminum production (and alumina con-
sumption) values of this calculation. All the other values like carbon con-
sumption or carbon dioxide, monoxide production are NOT relevant for the
aluminum chloride electrolysis.
about these calculations see Chap. 18.3.1

2-31
2. Mass Balance
Exercise 2-5:
Anode Consumption in 1 MAh
The anode gross consumption is AGC = 0.520. How much anode is con-
sumed in one mega ampere hour (CE = 92.4 %)?
The specific gross anode consumption (AGC) is given by the anode con-
sumption (AC) over the electrolytic aluminum production (EPAl, Equation
2-16)
AC
AGC =
EPAl
For 1 MAh one can write
AC = 0.520 × 0.3356 × 1000 × 0.924 = 161 kg

2-32
2. Mass Balance
Exercise 2-6:
Sodium Oxide in Alumina
Alumina contains about 0.5 wt % sodium oxide. Sodium oxide reacts with
aluminum fluoride according to
3 Na2O + 4 AlF3 = 2 Na3AlF6 + Al2O3
A pot is working at I = 200 kA and CE = 94 %. How much aluminum fluo-

ride must be added to compensate theoretically for this sodium oxide addi-
tion and what is the corresponding specific aluminum fluoride consump-
tion?
According to Table 2-2 185.9370 g of Na2O react with 335.9068 g AlF3.

Using Equation 2-16 and the specific electrolytic consumption of alumina
(Equation 2-19) one can write for the daily consumption of aluminum fluo-
ride
335.9
C AlF = 0.3356 ¥ 200 ¥ 0.94 ¥ 24 ¥ 189
. ¥ 0.005 ¥ = 25.9 kg
3
185.9
The specific consumption of aluminum fluoride is
335.9
SC AlF = 189
. × 0.005 × = 17 g / kg .
3
185.9

2-33
2. Mass Balance
Exercise 2-7:
Reaction of Sodium Carbonate with Aluminum Fluoride
The aluminum fluoride concentration with 15 wt % is too high. You want to

reduce it to 11 wt % (bath mass 3 t). To calculate how much sodium car-
bonate must be added write the reaction equation (sodium carbonate and
aluminum fluoride react to sodium cryolite, alumina and carbon dioxide).
The reaction equation is written as
3 Na2CO3 + 4 AlF3 = 2 Na3AlF6 + Al2O3 + 3 CO2

meaning that 317.9667 g sodium carbonate and 335.9068 g aluminum
fluoride react to 419.8226 g sodium cryolite, 101.9612 g alumina and
132.0297 g carbon dioxide. To remove 1 g AlF3 0.9463 g Na2CO3 are
needed.
You write for the mass of Na2CO3 needed
15 - 11
mNa2CO 3 = ¥ 3000 ¥ 0.9463 = 113.6 kg
100

2-34
3. Properties of the
Electrolyte
3.1. Summary ...................................................................... 3-3
3.2. Introduction ................................................................. 3-3
3.3. Composition of the Electrolyte .................................. 3-4
3.4. Bath Analysis............................................................. 3-12
3.5. Phase Diagrams ........................................................ 3-15
3.6. Liquidus Temperature............................................... 3-20
3.7. Electrical Conductivity.............................................. 3-22
3.8. Density ....................................................................... 3-24
3.9. Alumina Solubility ..................................................... 3-28
3.10. Viscosity ................................................................. 3-29
3.11. Surface Properties ................................................. 3-32
3.12. Vapor Pressure....................................................... 3-33
3.13. Literature................................................................. 3-38
3.14. Exercises ................................................................ 3-40

3-1

3-2
3.1. Summary
The properties of the electrolyte which depend on its composition

and the bath temperature influence the results of pot operation. The
chapter gives equations to calculate these properties and examples
to show their applications.
The composition of the electrolyte is chemically characterized in
several ways: by ratio and concentration values. The chemical labo-
ratory of a plant determines as a matter of routine the aluminum fluo-
ride, calcium fluoride and eventually the aluminum oxide content of
the electrolyte. The methods of chemical analysis are described. A
section about phase diagrams describes the meaning of some phase
diagrams for aluminum production.
In the following essential properties of the electrolyte are treated like
liquidus temperature, electrical conductivity, density, viscosity and
surface properties.
3.2. Introduction
The properties of the cryolite electrolyte influence many key parameters
and processes of the Hall-Héroult electrolysis. Some examples:
density: magnetohydrodynamic effects,
electrical conductivity: bath voltage, energy consumption,
solidus, liquidus temperature: heat transfer, ledge profile,
solubility of alumina: formation and dissolution of sludge,
Viscosity: dissolution of alumina,

electromagnetic effects,
surface tension: wetting of carbon anodes, anode effect.
Most of the properties depend on the composition and the temperature of

the electrolyte. In the following relations are given for these properties and
also indications how to use them.

3-3
3.3. Composition of the Electrolyte

The composition of the cryolite electrolyte is described in several ways:
different systems: sodium fluoride-aluminum fluoride-lithium

fluoride (NaF - AlF3 - LiF)
sodium cryolite-lithium cryolite
(Na3AlF6 - Li3AlF6)
different concentration mole percent (c)
units:
weight percent (x)
ratio of masses or sodium fluoride-aluminum fluoride ratio
moles: mNaF nNaF
,
m AlF n AlF
3 3
As Standard System sodium cryolite with additions is chosen:
Na3AlF6 - AlF3 - CaF2 - Al2O3 - LiF - ... (3–1)
The compositions are given in weight percent (wt %).
Remark about using Percent:

According to the ISO Standards Handbook (Lit. 2-1, p. 14) and a pamphlet
of the National Institute of Standards and Technology (NIST, Lit. 2-2, p.
20) the symbol % (per cent) is used for the number 0.01. Since percent is
a number it is in principle meaningless to speak about percentage by
mass. The preferred way to express a mass or mole fraction should there-
fore be "the mass (or mole) fraction is 67 %". We continue however to use
the expressions weight percent (wt %) and mole percent (mol %) in the
same way we use the current efficiency in percent (Equation 2-11).
Often the concentrations of AlF3 and NaF are called Excess Aluminum
Fluoride or Excess Sodium Fluoride Concentration. This specification
is not necessary since the standard system is defined as a multicompo-
nent system with cryolite as one component and the additions as the other
components.
To convert composition from mole fractions into weight fractions (and vice
versa) please use Table 3-1. These relations may be easily extended to
systems of more components.

3-4
Table 3-1: Conversion Mole Fractions <=> Weight Fractions.
Example 3-1 shows how to convert mole percent into weight percent.
Example 3-1: Conversion of Mole Percent into Weight Percent.
The composition of the electrolyte is 30 mol % AlF3 and 70 mol % NaF.
What is its composition in weight percent?
Using the relative molecular masses from Table 2-1 and the conversion re-
lations from Table 3-1 you can write:
x AlF ⋅ M AlF 0.3 × 84
c AlF = 3 3
= = 0.46
3
xAlF3⋅M e
+ 1− x
AlF3 ⋅M
AlF3 j 0.3 × 84 + 0.7 × 42
NaF
b
c AlF = 46 wt % weight percent
3
g
b
c NaF = 100 − 46 = 54 wt % weight percent g
(Chap. 18.14) on Exp. 3.1 ElysePrg introduces the start composition of
Exp. 3-1 namely 30 mole percent aluminum fluoride:

3-5
Remark: Since the alumina concentration is zero ElysePrg indicates an

anode effect. An anode effect influences the output of other pages of Ely-
sePrg (for instance the Pot Voltage page).
If you change the Concentration Units to Mass Fractions ElysePrg shows

the result in weight percent aluminum fluoride.
These relations may be easily extended to systems of more components

as you can see in Example 3-2.
Example 3-2: Conversion of the Compositions of a Three Component System.
The composition of the electrolyte is 29.3 mol % AlF3 and 2.7 mol % Al2O3
(the rest is NaF). What is its composition in weight percent?
According to Example 3-1 the relations from Table 3-1 must be extended:

3-6
x AlF ⋅ M AlF
c AlF = 3 3
3
3 2 3 2 3
e 3 2 3
j
x AlF ⋅ M AlF + x Al O ⋅ M Al O + 1− x AlF − x Al O ⋅ M NaF
3
0.293 × 84
c AlF =
3 0.293 × 84 + 0.027 × 102 + 0.680 × 42
b
= 0.44 44 wt % g
x Al O ⋅ M Al O
c Al O = 2 3 2 3
2 3
3 2 3 2 3
e
x AlF ⋅ M AlF + x Al O ⋅ M Al O + 1− x AlF − x Al O ⋅ M NaF
3 3 2 3
j
0.027 × 102
c Al O =
2 3 0.293 × 84 + 0.027 × 102 + 0.680 × 42
= 0.05 5 wt % b g
c NaF = 100 − 44 − 5 = 51 wt %
(you find the relative molecular masses in Table 2-2).

(Chap. 18.14) on Exp. 3.2 introduces the start composition of Exp. 3-2
namely 29.31 mole percent aluminum fluoride and 2.74 mole percent alu-
minum oxide:
If you change the Concentration Units to Mass Fractions ElysePrg shows

the result in weight percent aluminum fluoride.
The molten salt system aluminum fluoride - sodium fluoride can be con-
sidered as a mixture of the Lewis acid AlF3 and the Lewis base NaF.

3-7
A Lewis Acid is defined as an electron pair acceptor and a Lewis Base

as an electron pair donor. The acid - base pair AlF3 - NaF is forming the
neutral compound cryolite. To take into consideration the acidity of the
molten electrolyte other values than concentrations are used to describe
the composition:
the Cryolite Ratio:
nNaF (3–2)
CR =
n AlF
3
and the Bath Ratio:
mNaF (3–3)
R=
m AlF
3
Writing:
nNaF mNaF ⋅ M AlF 83.9767 CR
3
CR = = =R × = 2R , R =
n AlF m AlF ⋅ M NaF 419882
. 2
3 3
the following relation holds:
CR (3–4)
R=
2
When other Lewis acids (MgF2) or Lewis bases (LiF) are present in the
electrolyte some authors modify the bath or cryolite ratio to the
Equivalent Weight Ratio (EWR):
Tabereaux (Lit. 3-1)

3-8
nNaF + nLiF − nMgF

EWR = 1 ⋅ 2 (3–5)
2 n AlF
3
Haupin (Lit. 3-2)

n + 0.32 × nLiF
EWR = 1 ⋅ NaF (3–6)
2 n AlF + 0.41× nMgF
3 2
Haupin (Lit. 3-2) is arguing that using ratios may be confusing to describe
bath properties, therefore concentrations should be applied.
Table 3-2 contains relations to convert bath ratio values into the aluminum
fluoride and sodium fluoride concentrations of the standard system (Equa-
tion 3-1) and the sodium fluoride - aluminum fluoride (NaF - AlF3 - addi-
tions) system.
Table 3-2: Conversion Concentration, Bath Ratio of Different Systems.
with the meaning of the symbols is:
Example 3-3 shows how to use Table 3-2 to convert the concentrations of
Example 3-1 to the standard system. Similarly the three component sys-
tem of Example 3-1 is transformed to the standard system.

3-9
Example 3-3: Transformation to the Standard System.

Convert (1.) the compositions of Example 3-1 (46 wt % AlF3, 54 wt % NaF)
and (2.) of Example 3-2 (44 wt % AlF3, 5 wt % Al2O3, rest NaF) into the
values of the standard system. Calculate also the corresponding bath ratio.
You write using the relations of Table 3-2:
3
5
3 3
2
3
a f 5
3
2
c AlF = ⋅ c AlF − ⋅ 1− A = ⋅ 0.46 − = 0.10
3
1− c AlF − A 1− 0.46
R= 3
= = 117
.
c AlF 0.46
3
The aluminum fluoride (excess) concentration of Example 3-1 is 10 wt %
(bath ratio 1.17).
c AlF =
3
5
⋅c
3 AlF3 3
2
a f 5 2
a
− ⋅ 1− A = ⋅ 0.46 − ⋅ 1− 0.05 = 0.10
3 3
f
1− c AlF − A 1− 0.44 − 0.05
R= 3
= = 116
.
c AlF 0.44
3
The aluminum fluoride (excess) concentration of example 3-2 is 10 wt %

(bath ratio 1.16).
(Chap. 18.14) on Exp. 3.3 and ElysePrg converts the concentration of
Example 3-2 into the standard system:

3-10
For quick conversion of concentrations into bath ratios (and vice versa)
please use Table 3-3.
Table 3-3: Values of the Bath Ratio and AlF3 Concentration

3-11
3.4. Bath Analysis

The two important variables which must be monitored in the electrolyte for
efficient operation are the content of aluminum fluoride and the content of
aluminum oxide. The aluminum fluoride concentration of the electrolyte
varies slowly but continuously due to the sodium oxide content of alumina,
to the absorbtion of bath constituents by the carbon bottom blocks, to the
vaporization of electrolyte and due to adsorbed species on alumina which
have passed the gas cleaning system. The alumina concentration changes
because the electrolytic process consumes alumina which is added batch-
wise or regularly by a feeding device.
Total Bath Analysis

The sensitivity of the X-Ray Fluorescence (XRF) Method has improved so
much that a complete determination of the elements (aluminum, sodium,
calcium, lithium, fluorine and oxygen) of a bath sample is possible. The
inter-element effects seem to be small and the calibration curves are
rather linear. Claims have been made that the method will give bath ratio
accuracies of ± 0.04 (Lit. 3-3 p.261).
Description of the X-Ray Fluorescence (XRF) Method:
When the atoms in a sample material are irradiated with high-energy primary x-ray pho-
tons, electrons are ejected as photoelectrons. This creates electron holes in one or more
of the orbitals, converting the atoms into ions which are unstable.
To restore the atoms to a more stable state, the holes of inner orbitals are filled with elec-
trons from outer orbitals. Such transition is accompanied by an energy emission of sec-
ondary x-ray photon - a phenomenon called Fluorescence.
The energy of emitted fluorescent photons is determined by the difference in energies

between the initial and final orbitals for the individual transitions which is characteristic for
the element. The information about the energy and intensity of the fluorescent radiation is
used to determine the concentration of the atoms e.g. element in the sample.
Determination of the Aluminum Fluoride Content

The Pyrotitration (Lit. 3-4) adds sodium fluoride to a molten bath sample
until all excess aluminum fluoride has reacted to sodium cryolite. To do so
the bath sample after adding NaCl and KCl is molten at 850 °C. A small
part of the melt is taken out with a tiny ring made of platinum wire. With an
indicator the acidity of the sample is tested. NaF tablets of known weight
are added to the melt and the test with the platinum ring is repeated until
the sample is neutral.

3-12
Other methods add a known amount of sodium fluoride directly to the

sample which is heated to 650 °C. Sodium fluoride reacts with aluminum
fluoride to cryolite. Sodium fluoride replaces also lithium fluoride in lithium
cryolite. The sintered material is then extracted with water which dissolves
the excess sodium fluoride and the lithium fluoride. The Thorium Nitrate
Method (Lit. 3-3 p.255) then determined the fluoride concentration in the
solution. The Sinter Method (Lit. 3-5) uses Atomic Absorption to measure
the sodium fluoride and lithium content in the solution.
Description of Atomic-Absorption (AA) Spectroscopy:
Atomic-Absorption (AA) spectroscopy uses the absorption of light to measure the con-
centration of gas-phase atoms. Since samples are usually liquids or solids, the atoms or
ions must be vaporized in a flame or in a graphite furnace. The light source is usually a
hollow cathode lamp of the element that is being measured. The atoms absorb ultraviolet
or visible light and make transitions to higher electronic levels. The concentration is de-
termined from the amount of absorption.
Other wet chemical methods like the titration with aluminum chloride or
with potassium hydroxide are described in (Lit. 3-3 p.254-256). These
methods tend to be laborious, time consuming and the results are often
not of the wanted precision and reproducibility.
The X-Ray Diffraction (XRD) Method is based on the proportionality be-
tween the intensity of a selected diffraction line for a component and the
concentration in which this component is present in the solid sample. The
method is more complex than simply measuring the diffraction peaks of
several compounds (Lit. 3-6). Aluminum fluoride and cryolite have reacted
to chiolite and calcium fluoride to calcium cryolite (Lit. 3-7). Normally the
concentrations of calcium and lithium fluoride are determined by X-Ray
Fluorescence and the content of aluminum fluoride is then derived from
the chiolite content determined with X-Ray Diffraction.

3-13
Description of the X-Ray Diffraction (XRD) Method (Lit. 3-8):
X-rays have wavelengths of the same order as the interatomic distances in crystalline
solids and are therefore scattered producing a Diffracted Secondary X-Ray Beam. To
measure the intensity of the diffracted X-rays the detector is moved on a circle around the
sample which is turned simultaneously by half the angle of the detector.
The diffraction pattern e.g. the X-ray intensity in function of the diffraction angle is charac-
teristic for the composition of the diffracting solid sample and is used to determine the
concentration of the components in the sample.
Determination of the Aluminum Oxide Content

An aqueous aluminum chloride solution dissolves all fluoride compounds
of the electrolyte samples. The washed and dried residue is weighed as-
suming to be aluminum oxide (Extraction Method). Discussing the Liquidus
Enigma, Tarcy et al. (Lit. 3-9) compared the extraction method with the
Carbothermal Reduction Method or LECO-Method. The LECO appara-
tus reacts the oxides of the electrolyte sample at high temperatures (about
3000 °C) to carbon monoxide and dioxide which are determined analyti-
cally. It was shown that the results of the extraction method deviate sys-
tematically from the LECO-method results. This measuring error of the ex-
traction method can best be eliminated by switching to a new analysis pro-
cedure or to change the sampling procedure or post sample treatment.

3-14
3.5. Phase Diagrams

A phase diagram shows the phases of a system which are present at a
given temperature, pressure and composition of the system. A Phase is
homogenous in its crystal structure and atomic arrangement and has the
same physical and chemical properties throughout. The phase is sur-
rounded by an interface and can be mechanically separated from its sur-
rounding .
Phase Diagram NaF - AlF3

The phase diagram sodium fluoride - aluminum fluoride (Lit. 3-10 Fig. 2-2,
p.40) is an example:
Figure 3-1: Phase Diagram of the System Sodium Fluoride - Aluminum Fluoride.
Lines separate the temperature - composition diagram into regions where
different phases are in equilibrium. You find the coexistent phases in Ta-
ble 3-4.

3-15
Table 3-4: Phases of the System NaF - AlF3 in Figure 3-1.
Heating the solid phases of (8, e.g. solid cryolite and solid sodium fluoride)
and (11, e.g. solid chiolite and solid aluminum fluoride) to the Eutectic
Temperatures (888 °C and 695 °C) they melt either to pure liquid (5) or to
a mixture of solid and liquid ( 6, e.g. solid cryolite and liquid, 7, e.g. solid
sodium fluoride and liquid and 9, e.g. solid chiolite and liquid, 12, e.g. solid
aluminum fluoride and melt).
When heating (9, e.g. solid chiolite and liquid, 10, e.g. solid cryolite and
chiolite) respectively cooling (4, e.g. solid cryolite and liquid) over the Peri-
tectic Temperature (734 °C) chiolite decomposed into cryolite and alumi-
num fluoride or chiolite is formed by the reaction of cryolite with aluminum
fluoride. The reaction which happens at the peritectic temperature is writ-
ten as:
5Na 3AlF6 + 4AlF3 ↔ 3Na 5Al3F14 (3–7)

3-16
Phase Diagram Na3AlF6 - AlF3

For practical purposes e.g. to use the standard system and weight per-
cents a part of the NaF - AlF3 phase diagram (see Figure 3-2)
.
Figure 3-2: Part of the phase diagram of the system NaF - AlF3 (Figure 3-1) which is
transformed to the phase diagram of the system Na3AlF6 - AlF3 (Figure 3-3).
is transformed to the phase diagram of the system Na3AlF6 - AlF3

(Lit. 3-11 Fig. 6, p. 25).
Figure 3-3: Phase diagram of the system Na3AlF6 - AlF3 (Lit. 3-11 Fig. 6, p.25). Contrary
to Figure 3-1 the units of the abscissa are weight percent of AlF3.
The coexistent phases are given in Table 3-5.

3-17
Table 3-5: Phases of the System Na3AlF6 - AlF3 in Figure 3-3.
Suppose a Na3AlF6 - AlF3 melt of 10 wt % AlF3 (90 % Na3AlF6) at 1000 °C

is cooled down (Figure 3-4 A). When the temperature of the melt reaches
the Liquidus Temperature (Figure 3-4 B) pure solid Na3AlF6 is precipitat-
ing.
Figure 3-4: Cooling a Cryolite-Aluminum Fluoride Melt.
Continuing this process e.g. cooling the melt cryolite solidifies constantly
and the composition of the melt follows the liquidus line e.g. the AlF3 con-
tent of the melt increases (Figure 3-4 C). When the temperature reaches
the peritectic value the whole melt solidifies and the Peritectic Reaction
(Equation 3-7) takes place (Figure 3-4 D).
These processes are essential when bath is solidifying or melting like the
formation and dissolution of side ledge and top crust. Normally the content
of cryolite side ledge and top crust (more basic e.g. less acidic) is in-
creased compared to the composition of the liquid bath.

3-18
Phase Diagram Na3AlF6 - Al2O3

The phase diagram of the system Na3AlF6 - Al2O3 (Lit. 3-3, Fig. 3.3, p. 48)
is also important for aluminum electrolysis.
Figure 3-5: Phase diagram of the system Na3AlF6 - Al2O3 (Lit. 3-3, Fig. 3.3, p. 48).
You find the coexistent phases in Table 3-6:

Table 3-6: Pases of the System Na3AlF6 - Al2O3 in Figure 3-5.
If a melt (1) the composition of which is on the left side of the eutectic (7)
e.g. cAl2O3 < 11 wt % is cooled, cryolite (2) is solidifying when the tempera-
ture of the melt is less than the Cryolite - Liquidus (6). If the composition of
the melt (1) is however on the right hand side of the eutectic (7) e.g. cAl2O3
> 11 wt %, solid aluminum oxide(3) is formed when the temperature of the
melt is less than the Alumina - Liquidus (5).
The alumina - liquidus determines the maximal solubility of alumina in
cryolite since it is the limit of alumina concentration. If the system enters
the phase region (3) either because the alumina concentration is increas-
ing or the temperature is decreasing solid alumina is precipitating. These

3-19
processes play an important role in the formation and dissolution of alu-

mina bottom crusts.
3.6. Liquidus Temperature

The liquidus temperature is the temperature at which solid material is pre-
cipitating when a liquid is cooled. The electrolyte is a multi-component sys-
tem of cryolite with additions of aluminum fluoride, calcium fluoride and
eventually lithium fluoride and magnesium fluoride. In the literature several
relations of the cryolite liquidus temperature for this multicomponent sys-
tem are found (see also the Cryolite Liquidus of the binary Na3AlF6 - AlF3
system, Figure 3-3). This document gives the most recent (1995) relation-
ship of Solheim (Lit. 3-12 Equation 4):
(3–8)
The meaning of the symbols is:
ElyseSem refers to several other equations of the literature. In Exam-

ple 3-4 the use of Equation 3-8 is shown. Example 3-4 in ElyseSem calcu-
lates in addition the liquidus temperatures according to the other equa-
tions.

3-20
Example 3-4: Determination of the Liquidus Temperature

The composition of the electrolyte (mass fraction) is: 13.3 % AlF3, 4.7 %
CaF2, 2.7 % Al2O3, 1.3 % LiF and 0.7 % MgF2. Calculate the cryolite liq-
uidus temperatures using the relationship of Solheim.

(Chap. 18.14) on Exp. 3-4 and ElysePrg calculates the liquidus tempera-
tures:
If one calculates the liquidus temperature of the electrolyte using its

chemical analysis and one of the shown relations, one finds that some
cells operate successfully for extended periods (weeks) at temperatures
below the calculated liquidus. Haupin (Lit. 3-13) has analyzed this so
called Liquidus Enigma.
He offers the following explanations:
(1) impurities not determined in the industrial bath analysis lower the true
liquidus,
(2) an analytical bias of bath composition may exist,
(3) a bias in how bath temperature is measured may exist,

3-21
(4) bath may remain supersaturated with alumina for extended period of
time,
(5) industrial cells may be operating with suspended precipitate.
(6) He considers explanation 1,3 and 5 as most likely.
Tarcy et al. (Lit. 3-9) have analyzed the analytical methods to determine
the bath composition especially concerning the aluminum oxide concentra-
tion. They have shown that the extraction method which is normally used
to determine the alumina content of bath samples delivers systematically
too low values. The observation of cells operating below the calculated liq-
uidus temperature is, to a large extend, accounted for a systematic alu-
mina measurement error. Some cases of liquidus enigma may also be due
to impurities or measurement errors especially of the bath temperature.
3.7. Electrical Conductivity

The electrical conductivity of the cryolite electrolyte is an important factor
in cell voltage and therefore in power efficiency. The unit of the electrical
(or better electrolytic Lit. 2-1 p. 178) conductivity is siemens per meter
(S/m or S/cm). Siemens (reciprocal ohm or eventually mho) is the unit of
electrical conductance.
The literature cites several equations, this document gives the most recent
(1994) relationship of Híveš ( Lit. 3-14, Equation 8):
(3–9)
The meaning of the symbols is:
Lit. 3-14 cites a second less precise equation (Equation 9). The equation is
easier to use for hand calculations because it contains mass fractions in-
stead of mole fractions:

3-22
(3–10)
with the meaning of the symbols is:
ElyseSem refers to several other equations of the literature. In Exam-

ple 3-5 the use of Equation 3-9 is shown. Example 3-5 in ElyseSem calcu-
lates in addition the electrical conductivity according to these other equa-
tions.
Example 3-5: Determination of the Electric Conductivitiy of the Electrolyte
The composition of the electrolyte (mass fraction) is the same as in Exam-
ple 3-4 namely: 13.3 % AlF3, 4.7 % CaF2, 2.7 % Al2O3, 1.3 % LiF and 0.7
% MgF2. Calculate the electric conductivity using Equation 3-10 of Híveš.
The bath temperature (948.1 °C) is the liquidus temperature plus 10 °C
(see Example 3-4):

3-23

(Chap. 18.14) on Exp. 3-5 and ElysePrg calculates the electric conductivi-
ties:
3.8. Density
The difference of the densities of the molten electrolyte and the molten
aluminum metal is important. In Lit. 3-21 (p. 4-126) and Lit. 3-15 (6.12 and
6.13 on p.102) you find equations for the Density of Liquid Aluminum in
function of the temperature:
Handbook (1997): ρ Al = 2.529 − 0.000233 ⋅ t

Gebhardt (1953): ρ Al = 2.615 − 0.0002648 ⋅ t (3–11)
Moormann (1921): ρ Al = 2.635 − 0.000272 ⋅ t
In the literature one finds several equations which calculate the Density of
Liquid Bath in function of the chemical composition and the temperature:

3-24
Solheim (1995 Lit. 3-19)
Silny (1995 Lit. 3-20)

3-25
Welch (Lit. 3-3, Equation 3.18 p.59)
Haupin (1987 Lit. 3-15, Equation 6.14, p.103)

3-26
Example 3-6 shows the application of ElysePrg to determine the density

of liquid aluminum and of liquid cryolite electrolyte using the different equa-
tions.
Example 3-6: Determination of the Density of Liquid Aluminum and Bath

% MgF2. Calculate the density of aluminum and of the bath using the dif-
ferent equations. Corresponding to Example 3-4 the bath temperature
(948.1 °C) is the liquidus temperature plus 10 °C:
(Chap. 18.14) on Exp. 3-6 and ElysePrg calculates the densities:

3-27
3.9. Alumina Solubility

The maximum aluminum oxide concentration of the cryolite electrolyte de-
pends on its composition and temperature (Lit. 3-16):
2
A = 119
. − 0.062 × c AlF − 0.0031 × c AlF
3 3
− 0.20 × c CaF
2
− 0.50 × c LiF
− 0.30 × c MgF
2
42 × c AlF × c LiF
3
+
2000 + c AlF × c LiF
3
B = 4.8 − 0.048 × c AlF

3
1. 5
2.2 × c LiF
+
3
10 + c LiF + 0.001 × c
AlF3
sat T F IB
c Al
2O3
= A ×
1000 H K (3–11)
Equation 3-13 is the relationship for the alumina liquidus of the multi com-
ponent system cryolite - aluminum oxide with additions like aluminum fluo-
ride, calcium fluoride, etc. solved for the alumina concentration.
Example 3-7 shows the application of the Equation 3-13 to determine the
maximum alumina concentration of the bath:
Example 3-7: Determination of the Alumina Solubilty:

% MgF2. Calculate the maximum alumina concentration. Corresponding to
Example 3-4 the bath temperature (948.1 °C) is the liquidus temperature
plus 10 °C:

3-28

(Chap. 18.14) on Exp. 3-7 and ElysePrg calculates the alumina solubilty:
3.10. Viscosity
All fluids possess a definite resistance to changes of form. This property, a
sort of internal friction is called Dynamic Viscosity. Kinematic Viscosity
is the ratio of dynamic viscosity to density. The unit of the dynamic viscos-
ity is Poise and of the kinematic viscosity Stoke.
Table 3-7 shows the units of viscosity:

3-29
Table 3-7: Units of Viscosity
Table 3-8 compares the viscosity values of different substances. The vis-
cosity at 1000 °C of liquid aluminum corresponds to the viscosity of water
and the viscosity of the electrolytic bath may be compared to the viscosi-
ties of mercury and liquid lead.
Table 3-8: Values of Viscosity
The viscosity of the electrolyte influences several hydrodynamic processes

in the cell: movement of metal droplets in the electrolyte, dissolution and
sedimentation of alumina particles, release of the gas bubbles from the
anode surface. A simplified equation for the Electrolyte Viscosity is given
in Lit. 3-3 Equation 3-19:
(3–12)

3-30
The Viscosity of Liquid Aluminum you find in Lit. 3-15 Equation 6.15:
η Al = 0.1492 × exp e 1984.5

T + 273.15
j (3–13)
Example 3-8 shows the application of the equations to determine the vis-
cosity of liquid aluminum and of liquid cryolite electrolyte.
Example 3-8: Determination of the Viscosity of Liquid Aluminum and Liquid Bath

% MgF2. Calculate the viscosity of liquid aluminum and of liquid bath using
the different equations. Corresponding to Example 3-4 the bath tempera-
ture (948.1 °C) is the liquidus temperature plus 10 °C:
Viscosity of Liquid Aluminum:
Viscosity of Liquid Electrolyte:

3-31

(Chap. 18.14) on Exp. 3-8 and ElysePrg calculates the viscosity:
3.11. Surface Properties

The surface properties cause strong effects on the electrolysis process.
The interfacial tension between the electrolyte and the carbon influences
the penetration of bath into the pores of the carbon anodes or carbon lin-
ing and the separation of carbon particles from the electrolyte. The interfa-
cial tension between molten aluminum and electrolyte may affect current
efficiency.
Local unbalance of surface tension that may result from surface active
agents at the interface can change the heat and mass transfer rates and
alter flow patterns. This so called Maragoni Effect seems to be important
especially since the difference in density between the electrolyte and the
molten aluminum is rather small (Lit. 3-17).
The wetting angle or contact angle is a measure of the degree of wetting
of a solid surface by a liquid phase and determines also the formation of
gas bubbles. It is determined by the interfacial tensions gas - solid, liquid -
solid and gas - liquid:
γ − γ LS
cosΘ = GS (3–14)
γ GL

3-32
Figure 3-6: Wetting Angle and Bubble Formation. At high temperature and alumina con-
centration in the electrolyte the wetting angle of liquid electrolyte on carbon is
small e.g. the bath wets well the carbon anode. Consequently small bubbles
which are easily detached are formed. At low temperature and low alumina
content the behavior is inverted: large gas bubbles stick to the anode surface
and the electrolyte wets the anode surface poorly.
Small wetting angles produce good wetting and small easily dislodged
bubbles. In contrary a liquid with large wetting angle forms droplets of liq-
uid on the solid surface and large gas bubbles which adhere on the solid
surface (Figure 3-6).
At the time being no relations for the surface tension or wetting angles
concerning the aluminum production could be found. From diagrams
(Lit. 2-5) one can deduce that for the wetting of carbon materials by cryo-
lite the wetting angle decreases with increasing bath temperature and in-
creasing alumina content in the electrolyte.
3.12. Vapor Pressure

W. Haupin and H. Kvande (Lit. 3-18) have treated the fluorine evolution of
Hall-Héroult electrolytic cells. This chapter is taking in essence the rela-
tions from this publication.
The relation for the Total Vapor Pressure is:
⎛ -A ⎞ (3–17)
pTotal = exp ⎜ + B⎟
⎝T ⎠
with:

3-33
A = 21011 − 12235 ⋅ R + 18862 ⋅ R 2 − 6310.5 ⋅ R 3 +

116.7 ⋅ cLiF − 55 ⋅ R ⋅ cLiF −
(3–17A)
151 ⋅ cMgF2 + 1.466 ⋅ cMgF2 −
2
6.7 ⋅ c Al2O3
B = 25.612 − 9.681 ⋅ R + 11.854 ⋅ R 2 − 3.8315 ⋅ R 3 + (3–17B)

0.025 ⋅ cLiF − 0.013 ⋅ R ⋅ cLiF −
0.0008 ⋅ cCaF −
2
0.08696 ⋅ cMgF2 + 0.001112 ⋅ cMgF2 −

2
0.11 ⋅ c Al2O3
1 + 0.193 ⋅ c Al2O3
The vapor contains as major species sodium fluoride (NaF), NaAlF4 and
the dimer Na2Al2F8.
The Partial Vapor Pressure of Sodium Fluoride is given with:
⎛ -182 ⎞
pNaF = pTotal ⋅ ⎜ + 0.2073 ⎟ ⋅
⎝ T ⎠
(3–18)
( −0.6366 + 1.449 ⋅ R − 1.068 ⋅ R 2 + 0.2556 ⋅R 3 )
For the Dimerisation Reaction
Na2 Al2F8 = 2NaAlF4 (3–19)
the Equilibrium Constant is written with:

3-34
pM2 ⎛ -21085 ⎞ (3–20)

Kp = = exp ⎜ + 15.45 ⎟
pD ⎝ T ⎠
Solving the quadratic equation for the monomer species gives the partial
pressures of the monomer and dimer gas species.
Partial Vapor Pressure of NaAlF4:
−1+ 1+ 4 ⋅ Kp ⋅ ( pTotal − pNaF ) (3–21)

pM =
2 ⋅ Kp
Partial Vapor Pressure of Na2Al2F8:
pD = pTotal − pNaF − pM (3–22)

Example 3-9 shows the application of the equations to determine the total
vapor pressure and the partial pressures of the gaseous species.
Example 3-9: Determination of the Vapor Pressures

% MgF2. Calculate the total vapor pressure and the partial pressures of the
gaseous species. Corresponding to Example 3-4 the bath temperature
(948.1 °C) is the liquidus temperature plus 10 °C:
The following equations use the bath ratio (see Exp. 3-5).
You determine the total vapor pressure according to the equations 3-17A,
3-17B and 3-17:

3-35
For the partial vapor pressure of sodium fluoride you write according to
equations 3-18:
and for the equilibrium constant of the dimerisation reaction (Equ. 3-20):
You find the partial vapor pressure of NaAlF4 (Equ. 3-21):
and of Na2Al2F8 (Equ. 3-22):

(Chap. 18.14) on Exp. 3-9 ElysePrg calculates the vapor pressures:

3-36

3-37
3.13. Literature
Lit. 3-1: A. T. Tabereaux, T. R. Alcorn , L. Trembley,
"Lithium-Modified Low Ratio Electrolyte Chemistry for Improved Per-
formance in Modern Reduction Cells",
Lit. 3-2: W. Haupin,

"Defining Bath Ratio in the Presence of LiF and MgF2",
Lit. 3-3: K. Grjotheim, B. J. Welch,

"Aluminum Smelter Technology",
2nd Edition,
Aluminium-Verlag, Düsseldorf, (1988).
ISBN 3-37017-162-6.
Lit. 3-4: Analysenvorschrift,

"Pyrotitration von Ofenfluss",
Alusuisse Internal Report.
Lit. 3-5: Analysenvorschrift,

"Sintermethode (AlF3), Extraktionsmethode (Al2O3)",
Alusuisse Internal Report.
Lit. 3-6: U. Mannweiler, W. Schmidt-Hatting, M. Koutny,

"Neuartige automatische Analyse der Badkomponenten der Schmelz-
flusselektrolyse von Tonerde",
Erzmetall (1983), 36, Nr. 6, 274-277.
Lit. 3-7: O. Herstad, C. Krohn, M. Sørlie, H. A. Øye,

"Precipitation of Solid Compounds during Electrolysis of Cryolite-
Alumina Melts",
Aluminium (1983), 59. Jahrg. (3), 200-206.
Lit. 3-8: Siemens AG,

"Neues Analysensystem für die Prozessführung der Aluminiumindu-
strie",
Analysentechnische Mitteilung Nr. 294 (1983),
Analysentechnik Siemens AG.
Lit. 3-9: G. P. Tarcy, S. Rolseth, J. Thonstad,

"Systematic Alumina Measurement Errors and Their Significance in
the Liquidus Enigma",
Lit. 3-10: K. Grjotheim, H. Kvande (Editors),

"Introduction to Aluminium Electrolysis",
2nd Edition,
ISBN 3-87017-233-9.
Lit. 3-11: M. Rolin,

"La Physicochimie des Bains d'Electrolyse de l'Aluminium",

3-38
Publication de l'Institut des Sciences Appliquées

de Lyon Villeurbanne, (1975).
Lit. 3-12: A. Solheim, S. Rolseth, E. Skybakmoen, L. Støen, Å. Sterten, T. Støre,

"Liquidus Temperature and Alumina Solubility in the System Na3AlF6-
AlF3-LiF-CaF2-MgF2",
Lit. 3-13: W. Haupin ,

"The Liquidus Enigma",
Lit. 3-14: J. Híveš, J. Thonstad, Å. Sterten, P. Fellner,

"Electrical Conductivity of Molten Cryolite-Based Mixtures Obtained
with a Tube-Type Cell made of Pyrolytic Boron Nitride",
Lit. 3-15: A. R. Burkin (Editor),

"Production of Aluminum and Alumina",
Critical Reports on Applied Chemistry Volume 20,
John Wiley & Sons, (1987).
ISBN 0 471 91424 X.
Lit. 3-16: E. Skybakmoen, A. Solheim, A. Sterten,

"Alumina Solubility in Molten Salt Systems of Interest for Alumina
Electrolysis and Related Phase Diagram Data ",
Met.Trans.B 1997, Vol. 28B, 81-86.
Lit. 3-17: T. A. Utigard, J. M. Toguri,

"Maragoni Flow in the Hall-Héroult Cell",
Lit. 3-18: W. Haupin, H. Kvande,

"Mathematical Model of Fluoride Evolution from Hall-Héroult Cells",
Lit. 3-19: H. Gudbrandsen,

"Systemizing the Properties of Hall-Héroult Baths",
Prospectus of Electrolysis Group, Process Metallurgy and Ceramics,
SINTEF Materials Technology.
Lit. 3-20: A. Silny,

"Physico-Chemical Properties of Aluminum Electrolytes",
Prospectus of Institute of Inorganic Chemistry, Slovak Academy of Science,
842 36 Bratislava, Slovak Republic
Lit. 3-21: D. R. Lide (Editor-in-Chief),

"Chemical Rubber Company Handbook of Chemistry and Physics",
77th Edition (1996-97).
CRC Press, Boca Raton, New York, London, Tokyo;

3-39
3.14. Exercises
3-1: Conversion of the Bath Ratio into Weight Percent.
3-2: Phase Diagram Na3AlF6 -CaF2.
3-3: Properties of Cryolite.
3-4: Change of Aluminum Fluoride Concentration.

3-40
Exercise 3-1:
Conversion of the Bath Ratio into Weight Percent:
The bath ratio of the electrolyte is R = 1.6 (no other additions). What is the
composition of the standard system?
Using Table 3-2 there are two solutions possibel:

If you use the the relationship for the aluminum fluoride concentration you
find:
2 2
1− ⋅R 1− × 16
.
3 3
c AlF = = = −0.026
3
1+ R 1 + 16
.
A negative value of the aluminum fluoride concentration means that the
electrolyte contains an excess of sodium fluoride e.g. the composition is
on the left side of cryolite in Figure 3-1 (phase field 6). Continuing to use
Table 3-2 you write:
3
cNaF = − ⋅ c AlF = −1.5 × ( −0.026 ) = 0.039
3
2
cNaF = 3.9 wt%
If you know however that R > 1.5 means an excess of sodium fluoride you
use the relationship for the sodium fluoride concentration in Table 3-2:
3
R −
2 . − 15
16 .
c NaF = = = 0.039
1+ R 1 + 16
.
c NaF = 3.9 wt%

3-41
Exercise 3-2:
Phase Diagram Na3AlF6 -CaF2:
In the literature you find the phase diagram cryolite - calcium fluoride. Indi-
cate the coexistent phases. Determine the composition of the eutectic.
Reading from the phase diagram that the eutectic is at about 50 mol% you
write using Table 2-2 for the molecular weights and Table 3-1 for the con-
version formulas:

3-42
Exercise 3-3:
Properties of Cryolite:
What are the properties of pure cryolite (use ElysePrg)?
You set in the bath property page of ElysePrg the concentrations of all the
additions and of the overheat to zero. Then you read:

3-43
Exercise 3-4:
Change of Aluminum Fluoride Concentration:
The ledge of an electrolytic cell (length = 7m, width = 2.5 m) is formed

along its vertical sides. The ledge (thickness = 10 cm, height = 30 cm)
consists of pure cryolite (density = 2.85 g/cm3). Estimate the relative
change of the aluminum fluoride concentration after completely melting the
ledge (mass of electrolyte = 3 t).
You write for the ledge mass:
m Ledge = 2 ⋅ ( l + w − 2 x ) ⋅ x ⋅ h ⋅ ρ Na AlF6
3
2 × ( 700 + 250 − 20 ) × 10 × 30 × 2.85

m Ledge = = 16
. t
6
10
with
You write for the aluminum fluoride concentration before melting of the
ledge:
m AlF
3
c1 =
m Bath
and after the melting:
m AlF
3
c2 =
m Bath + m Ledge

3-44
and for the relative change of the aluminum fluoride concentration:
Suppose the aluminum fluoride concentration is 12 % it would be reduced

to about 8 % after melting of the entire ledge.

3-45

3-46
4. Pot Voltage
4. Pot Voltage
4.1. Summary ...................................................................... 4-3
4.2. Components of Pot Voltage ....................................... 4-3
4.3. Components of Bath Voltage ..................................... 4-5
4.4. Reversible Decomposition Voltage ........................... 4-6
4.5. Ohmic Bath Voltage .................................................... 4-8
4.6. Bubble Voltage Drop................................................. 4-12
4.7. Overvoltages ............................................................. 4-13
4.8. Bath Voltage and Alumina Concentration .............. 4-14
4.9. Anode Effect .............................................................. 4-16
4.10. Literature................................................................. 4-17
4.11. Exercises ................................................................ 4-18

4-1
4. Pot Voltage

4-2
4. Pot Voltage
4. Pot Voltage
4.1. Summary
The different components of the pot voltage inside and outside of the
electrolysis cell are described.
The bath voltage contributes the greatest part to the pot voltage. Re-
lationships for its ohmic and electrochemical parts are given namely:
the reversible decomposition voltage, the ohmic bath voltage, the
bubble voltage drop and the overvoltages.
The bath voltage depends on the alumina concentration in the elec-
trolytic bath. This relationship is the base of the automatic control of
the alumina concentration by computer process control.
When the alumina concentration is low the electrolysis reactions
change and the anode effect occurs. The mechanism of the anode
effect and its quenching is described.
4.2. Components of Pot Voltage

The Pot Voltage is the entire voltage applied to an electrolysis pot. The
pot voltmeter shows this value which is measured from the entrance to the
exit of the pot busbar system.
Figure 4-1: Pot Voltage: the pot voltage is the entire voltage applied to an electrolysis
pot. It is measured from the entrance to the exit of the busbar system.
The pot voltage is divided into several components:

4-3
4. Pot Voltage
Figure 4-2: Components of the Pot Voltage: Entry to Anode Voltage, Bath Voltage, Bot-
tom (Cathode) and Collector Bars to Exit Voltage.
Entry to Anode Voltage (voltage drop from input to anode, Figure 4-3):
the value of the anode voltage checks the electrical connections between
the different parts of the anode system (carbon anodes, iron yoke, alumi-
num rod, anode clamp, and anode beam).
Figure 4-3: Entry to Anode Voltage: the Entry to Anode Voltage Drop contains the volt-
age from the entrance to the anode.
Bath Voltage (Figure 4-4): the bath voltage contains a resistance and an
electrochemical part. The measured value of the bath voltage determines
the set value of the pot voltage.
Figure 4-4: Bath Voltage: the bath voltage contains a resistance and an electrochemical
part.
Cathode Voltage or Bottom Voltage (Figure 4-5): the cathode voltage or

bottom voltage is determined by the amount of bottom sludge or bottom
crust, the quality and the ageing of the carbon bottom blocks.
4-4
4. Pot Voltage
Figure 4-5: Cathode Voltage or Bottom Voltage: is the voltage drop from the bottom lin-
ing to the cathode bus.
Collectro Bars to Exit Voltage (Figure 4-6): the external voltage is the
voltage drop in the external bus bar system.
Figure 4-6: Collectro Bars to Exit Voltage: the external voltage is the voltage drop in the
external bus bar system
The Figures 4-1 to 4-6 show the definition of the different voltage drops.
They do not indicate, however, how these voltage drops are actually de-
termined e.g. measured practically during pot operation (see Chapter 14).
4.3. Components of Bath Voltage

The bath voltage contributes with 3.5-4.0 V the greatest part to the pot
voltage. The bath voltage is measured in the anode to cathode distance
(ACD) e.g. between the carbon anodes and the aluminum metal cathode.
Figure 4-6 shows the components of the bath voltage:
Figure 4-7: Components of the Bath Voltage: the bath voltage is the sum of resistance
components (UΩ: ohmic bath voltage, UBub: bubble voltage) and electro-

4-5
4. Pot Voltage
chemical components (E0: reversible electromotoric force, ηAC,ηAR: anodic

concentration and reaction overvoltage, ηCC: cathodic concentration over-
voltage).
E0 Reversible Electromotoric Force: the minimum voltage which is nec-

essary to produce aluminum according to the Principal Equation
(Equation 2-1) in an ideal electrolysis cell.
UΩ Ohmic Bath Voltage: the voltage loss due to the ohmic resistance of
the electrolyte.
UBub Bubble Voltage: according to the Electrolysis Equation (Equa-

tion 2-13) the anodes produce gas which forms a layer in the anode
to cathode distance. The layer is an insulator for the electrolysis
current. It increases the resistance of the electrolyte e.g. the bath
voltage.
ηAC Anodic Concentration Overvoltage: this voltage is necessary to

overcome the concentration gradient of the reacting species at the
anode.
ηAR Anodic Reaction Overvoltage: an extra voltage has to be applied to

make the reaction at the anode proceed with an appropriate speed.
ηCC Cathodic Concentration Overvoltage: this voltage is necessary to

overcome the concentration gradient of the reacting species at the
metal pad cathode.
4.4. Reversible Decomposition Voltage

Electrochemistry uses the concept of an ideal electrochemical cell. When
this cell is in equilibrium, i.e. no electric current is flowing in the system,
the measured voltage is called the reversible electromotive force (EMF).
Applied to the Principal Equation (Equation 2-1) of the Hall-Héroult-
Process this reversible electromotive force or reversible decomposition
voltage is the minimum voltage to produce aluminum in an ideal electroly-
sis cell (see also Chapter 5-5). In reality, however, when an electric current
is flowing through the pot and aluminum is produced additional voltages
have to be applied:
Ohmic Voltage Drop: this voltage is necessary to overcome the electrical

resistance in the busbar system and in the electrolytic cell.
Overvoltages: due to the electrochemical processes on the electrodes
additional electrical energy i.e. voltage is necessary to transport the
electroactive species to the electrodes and to keep them reacting at
a reasonable speed.

4-6
4. Pot Voltage
The Reversible Decomposition Voltage E0 (Lit. 4-2) is mostly related to

the bath temperature, alumina concentration and alumina solubilty (see
Chap. 3-8):
T + 273.15 T + 273.15 c Al O (4-1)

E 0 = 18932
. − − ⋅ ln sat2 3
1757.7 25138 c Al2 O 3
Example 4-1 determines the reversible decomposition voltage for an

Alusuisse EPT18 pot:
Example 4-1: Determination of the Reversible Decomposition Voltage
Calculate the reversible decomposition voltage for the Alusuisse 180 kA
pot. The composition of the electrolyte (mass fraction) is: 12 % AlF3, 5 %
CaF2, 2.5 % Al2O3, 0.5 % LiF and 0.5 % MgF2. Use ElysePrg to calculate
the bath temperature with an overheat of 10 °C and with the equation of
Solheim for the liquidus temperature. Use ElysePrg as well to determine
the alumina solubility.
ElysePrg determined a bath temperature of 962.2 °C and alumina solubility

of 8 wt%. Using Equation 4-1 one writes:

(Chap. 18.14) on Exp. 4-1 and ElysePrg executes the calculation of Exp.
4-1. You find the Reversible Decomposition Voltage in the Electrochemical
Bath Voltage window:

4-7
4. Pot Voltage
4.5. Ohmic Bath Voltage

During electrolysis the newly set prebaked carbon anodes are consumed
e.g. the sharp corners are gradually rounded off. The anode attains a
steady state shape after six to nine days.
Figure 4-8: Fanning of the Electrolysis Current: the electrolysis reactions consume the
sharp corners of the carbon anodes. The ideal current distribution is
changed (ACD: anode to cathode distance, AA: surface of new anodes). The
anodic (fA) and the cathodic (fC)fanning factors take care of this surface
change.
In order to calculate the electrical resistance e.g. the ohmic voltage drop
across the bath, the calculation account for these variations in current
density and cross sectional areas by using the so called Anodic (fA) and
Cathodic Fanning Factor (fc). The Anodic (jA) and the Cathodic (jC)
Current Density is then written:
I I (4-2)
jA = , jC =
f A ⋅ AA f C ⋅ AA

4-8
4. Pot Voltage
The Anode Surface (AA) is calculated by assuming that all prebaked an-
odes of a pot are new:
AA = n A ⋅ l A ⋅ w A (4-3)
Introducing into Ohm's Law:
U = I ⋅R (4-4)
the electrical resistance of the electrolyte:
1 d 1 ACD (4-5)
R= ⋅ = ⋅
κ A κ f A ⋅ AA
an expression for the Ohmic Voltage Drop (UΩ) of the bath is found:
ACD ⋅ j A (4-6)
UΩ =
κ

4-9
4. Pot Voltage
In chapter 3.6 electrical conductivity equations are given to calculate the

electrical conductivity of the electrolyte in dependence of its chemical
composition and temperature.
Example 4-2 determines the ohmic bath voltage drop for different pot
types.
Example 4-2: Determination of the Ohmic Bath Voltage
Estimate the ohmic bath voltage of the Alusuisse 180 kA pot (Table 1-1), of
Pechiney 300 kA pot (Table 1-3) and of VAW 300 kA pot (Table 1-4). Use
the composition of the electrolyte of Example 4-1 (5 % CaF2, 2.5 % Al2O3,
0.5 % LiF and 0.5 % MgF2). Use ElysePrg to calculate the bath tempera-
ture with an overheat of 10 °C and with the equation of Solheim for the liq-
uidus temperature. Use ElysePrg as well to calculate the electrical conduc-
tivity of the electrolyte with the relation of Híveš. The anodic fanning factor
is set to fA = 1.05, the anode to cathode distance to ACD = 5 cm and the
AlF3 concentration to cAlF3 = 12 % for Alusuisse's 180 kA pot.
Alusuisse:
With an AlF3 concentration of 12 % ElysePrg determined a bath tempera-
ture of 962.2 °C and an electrical conductivity of 2.1635 S/cm. The anodic
current density and ohmic bath voltage is then calculated using Equa-
tion 4-1 with the values of Table 1-1 and Equation 4-5.

(Chap. 18.14) on Exp. 4-2-1 and ElysePrg executes the calculation for the
Alusuisse pot. You find the voltage drop in the Ohmic Bath Voltage win-
dow:
Since the bubble voltage is zero the voltage drop and the ohmic bath volt-
age are equal.
Pechiney:
AlF3 concentration: 11 % (Table 1-3), Bath temperature: 966.8 °C and
electrical conductivity: 2.1992 S/cm.

4-10
4. Pot Voltage

Pechiney pot. You find the voltage drop in the Ohmic Bath Voltage window:
age are equal.
VAW:
AlF3 concentration: 8.5 % (Table 1-4), Bath temperature: 975.7 °C and
electrical conductivity: 2.2856 S/cm.

VAW pot. You find the voltage drop in the Ohmic Bath Voltage window:
age are equal.

4-11
4. Pot Voltage
4.6. Bubble Voltage Drop

The horizontal orientation of the anode in the Hall-Héroult-cell causes gas
accumulation beneath the surface, reducing the electrically conducting
area. This increases the effective resistivity of the electrolyte with the re-
sulting increase of the bath voltage referred to as bubble voltage drop.
Lit. 4-1 reviews the different bubble resistance models and introduces the
following equation:
d ⋅j ϕ
U Bub = Bub A ⋅ (4-7)
κ a1− ϕ f
The values of the bubble layer thicknes and of the surface coverage are in
the order of (Lit. 4-2):
dBub = 0.55 - 0.45 cm, ϕ = 0.50 - 0.90,

for an anodic current density of jA = 0.25 - 1.00 Acm-2.
Example 4-3: determines the bubble voltage drop for the Alusuisse EPT18
pot:
Example 4-3: Determination of the Bubble Voltage Drop
Estimate the bubble voltage drop of the Alusuisse 180 kA pot assuming a
bubble layer thickness of 0.5 cm and a surface coverage of 0.8. Use the
values (anodic current density and electrical conductivity) of Example 4-2.
Using Equation 4-6 one writes:

(Chap. 18.14) on Exp. 4-3 and ElysePrg executes the calculation for the
Alusuisse pot. You find the bubble voltage in the Ohmic Bath Voltage win-
dow:

4-12
4. Pot Voltage
Since the bubble voltage is not zero the ohmic bath voltage is the sum of
the voltage drop and and the bubble voltage.
4.7. Overvoltages
In addition to the reversible decomposition voltage and to the voltages to
overcome the electrical resistance a supplementary voltage must be ap-
plied to execute the electrolysis process. Electrochemistry distinguishes
two types of overvoltage (Lit. 4-3):
Charge Transfer Overvoltage (Reaction Overvoltage): the excess ap-

plied voltage makes the reactions especially the electron transfer at
the electrodes proceed at an appropriate speed.
Concentration Overvoltage (Polarization): due to the slow mass trans-

port of the reacting species a concentration gradient is built up
around the electrodes. The additional concentration overvoltage is
necessary to make the electroactive species migrate against this
concentration gradient.
In Lit. 4-3 p. 144 the following semi empirical relations are given:
Reaction Overvoltage on the Anode:
η AR =
aT + 273.15f .lnFG j A IJ
12533 H jlimit K (4-8)

4-13
4. Pot Voltage
Concentration Overvoltage on the Anode:
η AC =
aT + 273.15f .lnFG j crit IJ
23209 H j crit − j A K (4-9)
where the limiting current density is given by:
0.56 (4-10)
j limit = 0.0029 ⋅ c Al O 2 3
and the critical current density by:
a 0.5
5.5 + 0.018 ⋅ T − 1323 ⋅ c Al f FH
2O 3
− 0.4 IK (4-11)
j crit = 0.1
AA
Concentration Overvoltage on the Cathode:
aT + 273.15f ⋅ a1375
. − 0.125 ⋅ CR f F j C I
.ln G
η CC =
17407 H 0.257JK (4-12)
4.8. Bath Voltage and Alumina Concentration

The bath voltage changes its value with the alumina content in the electro-
lyte. In Figure 4-9 the dependence of the ohmic part (ohmic bath voltage
and bubble voltage) and the electrochemic part (reversible decomposition
voltage and anodic and cathodic overvoltages) on the alumina concentra-
tion is shown.

4-14
4. Pot Voltage
Figure 4-9: Ohmic (UΩ empty circles) and electrolytic (Uel crosses) parts of the bath volt-
age in dependence of the alumina concentration.
The ohmic part increases because in essence the electrical conductivity of
the electrolyte decreases with increasing alumina concentration. Contrary
the electrochemic part decreases with increasing alumina concentration.
One can imagine that the higher the alumina concentration the less energy
is necessary to convert aluminum oxide to aluminum.
Summing up UΩ and Uel give the global bath voltage (UBath) which can be
measured directly with the volt meter (Figure 4-10).
Figure 4-10: Bath voltage (UBath) in dependence of the alumina concentration.

The curve bath voltage vs. alumina concentration is the basis of the auto-
matic control of the alumina concentration by the computer process con-
trol. The process control changes the alumina concentration in the electro-
lyte by modifying the alumina feeding rate (over or under feeeding). Ana-
lyzing the voltage behaviour with time the process control algorithm de-
termines the alumina concentration and prescribes the optimal alumina
feeding rate.
The curves of Figure 4-9 and Figure 4-10 were determined with ElysePrg
(see the exercise). They should show the principle and give values which
are smaller than observed on an electrolytic pot. One of the reasons is that

4-15
4. Pot Voltage
the bubble layer parameters were kept constant e.g. the influence of the
alumina concentration on the bubble layer thickness and on the surface
coverage was neglected.
4.9. Anode Effect

The anode effect occurs in an electrolytic pot when the content of dis-
solved alumina in the bath becomes too low e.g. about less than 1.5 wt %
of alumina. During the anode effect the cell voltage increases rapidly,
since the electric current intensity is kept constant in a line of cells con-
nected in series. The voltage may vary considerably, to 10 to 50 volts de-
pending on the operating conditions of the cell, particularly of the anodes.
Arcs may be observed on the anode surface and a crackling can be heard.
Carbon tetrafluoride is evolved with a minor amount of other fluorides.
For a given alumina concentration and bath temperature the onset of the
anode effect is rather reproducible and occurs at the so called critical cur-
rent density. The critical current density depends on the alumina concen-
tration and on the temperature: lowering the alumina concentration and/or
the temperature decreases the critical current density.
Anode effects were thought to be necessary for pot operation as an ana-
lytical means because when an anode effect occurred all the added alu-
mina was consumed and the cell was not overfed and for its cleansing ac-
tion to bring carbon dust on the bath surface and burn it there. With im-
proved quality control of anode manufacture and more automated cell op-
eration these arguments are becoming less valid. Consequently cells have
been operated for several weeks without anode effects and negative ef-
fects on cell performance.
Mechanism of the Anode Effect
During anode effect the electrolysis mechanism is changed: the electrolyte

seems to be also electrolysed. Since this should only increase the voltage
by about one volt other additional mechanism were proposed:
• insulating layer of solid material (electrolyte) is formed on the anode,

• the gas bubbles are electrostatically attracted to the anode,
• an insulating gas film is formed due to surface effects,
• kinetic phenomena.
Visual observation of the anode effect supports the gas film explanation.
Quenching of the Anode Effect

To terminate the anode effect it is not sufficient to increase the alumina
concentration in the bath by overfeeding. The gas film under the anodes
must be removed by mechanical action. Most modern cells have automatic
anode effect termination which replaces the green wooden poles which
were inserted manually into the bath under the anodes. The anode table is
lowered by process control for a short time (a few seconds) and moved up
4-16
4. Pot Voltage
again in several cycles. Another method is to tilt the anode table avoiding
the overflow of electrolyte. It was also discussed to short circuit some an-
odes or blow air through pipes under the anodes .
4.10. Literature
Lit. 4-1: T. M. Hyde, B. J. Welch,

"The Gas under Anodes in Aluminium Smelting Cells Part I: Measuring
and Modelling Bubble Resistance under Horizontally Oriented Elec-
trodes",
Lit. 4-2: R. J. Aaberg, V. Ranum, K. Williamson, B. J. Welch,

"The Gas under Anodes in Aluminium Smelting Cells Part II: Gas Vol-
ume and Bubble Layer Characteristics",

2nd Edition,
ISBN 3-87017-233-9.

4-17
4. Pot Voltage
4.11. Exercises
4-1: Bath Voltage vs. Alumina Concentration.
4-2: Electrolysis on Carbon Particle.

4-18
4. Pot Voltage
Exercise 4-1:
Bath Voltage vs. Alumina Concentration:

Determine the bath voltage in dependence of the alumina concentration
(from 1.5 to 6.5 wt% in steps of 1 %). Use ElysePrg with the following pa-
rameters:
Bath composition: 13.3 % AlF3, 4.7 % CaF2, 1.3 % LiF, 0.7 % MgF2.
Overheat: constant 10 °C.
Current Intensity: 185 kA.
32 Anodes, length: 1450 mm, width: 540 mm.
Anode to Cathode Distance: 4.5 cm.
Bubble Layer Thickness: 0.5 cm.
Surface Coverage: 0.5.
Anodic Fanning Factor: 1.1.
Cathodic Fanning Factor: 1.2.
Add the ohmic bath voltage and bubble voltage for the ohmic part and the
reversible decomposition voltage and the anodic and cathodic overvolt-
ages for the electrochemic part.
On the bath properties page of ElysePrg you fill in the bath composition
starting with an alumina concentration of 1.5 %. Check constant overheat
and put in the value of 10 °C.
On the geometry page fill in the anode number and dimensions.
On the pot voltage page fill in the current intensity, check anode to cath-
ode distance fill in the value and fill in the values of bubble layer thickness,
surface coverage, anodic and cathodic fanning factor.
With the following relationships:

4-19
4. Pot Voltage
you calculate the first value for the bath voltage and its ohmic and electro-
chemic part.
Go back to the bath properties page increase the alumina concentration to
2.5 % and get again the voltage values from the voltage page.
After continuing to an alumina concentration of 6.5 % you get the following
table:

(Chap. 18.14) on Exs. 4-1 corresponding to the different alumina concen-
trations. If you press the line for cAl2O3 = 3.5%, for instance, you receive
the following result:

4-20
4. Pot Voltage
Exercise 4-2:
Electrolysis on Carbon Particle:

A carbon particle floats in the cryolite electrolyte. An electrolysis current is
flowing across this bath with a current density of 1 ampere per square cen-
timeter. What is the minimum diameter of this particle that electrolysis
takes place on this particle (electrochemic decomposition voltage: 1.65 V,
electrical conductivity of the electrolyte: 2.15 S/cm)?
Rewriting Equation 4-6 you find
The diameter of the carbon particle is:
1.65 × 2.15
d= = 3.5cm
1
During an anode effect the apparent electrical conductivity is considerably

reduced. Much smaller carbon particles (carbon dust) are then consumed
by this local electrolysis mechanism.

4-21
4. Pot Voltage

4-22
5. Energy Balance
5. Energy Balance
5.1. Summary................................................................... 5-3
5.2. Introduction .............................................................. 5-3
5.3. Energy Requirements .............................................. 5-4
5.4. Heat Loss .................................................................. 5-7
5.5. Reversible Decomposition Voltage ........................ 5-9
5.6. Literature................................................................. 5-11

5-1
5. Energy Balance

5-2
5. Energy Balance
5. Energy Balance
5.1. Summary
The relationship for the specific energy consumption is derived.

According to the First Law of Thermodynamics the amount of energy
is determined to produce aluminum from aluminum oxide. All that
energy is lost as heat which is not needed for the electrolytic produc-
tion of aluminum. Relationships are given to calculate the amount of
this heat loss.
Using the Nernst Equation i.e. the Second Law of Thermodynamics
the reversible decomposition voltage is derived. The reversible de-
composition voltage is the minimal voltage to produce electrolyti-
cally aluminum.
5.2. Introduction
The aluminum smelting process needs energy to keep the electrolytic cell
at the reaction temperature and to produce aluminum. This required en-
ergy is furnished as electric energy. The power (watts) which is added to a
pot is written with
P = U Pot ⋅ I (5–1)
The energy (E) to produce one kilogramm of aluminum (Specific Energy

Consumption) is derived in the following way by using Equ. 2-16 for the
electrolytic production of aluminum (EPAl):
P ⋅t U Pot ⋅ I ⋅ t
E = =
EPAl 0.3356 ⋅ I ⋅ t ⋅ η
U U
E= = 297.97 ⋅ (5–2)
0.3356 ⋅ η CE

5-3
5. Energy Balance
The difference between the added energy and the energy required to pro-
duce aluminum is lost as heat. This heat loss is of major concern since pot
operation wants to keep its value at a minimum without disturbing optimal
pot operation.
5.3. Energy Requirements

The First Law of Thermodynamics also called the Law of Conservation of
Energy states that, although energy may be converted from one form to
another, energy cannot be created or destroyed. Whenever a quantity of
one kind of energy disappears an exactly equivalent amount of the other
kind must be produced. Table 5-1 shows the different units of energy.
Table 5-1: Conversion of Energy Units
Thermodynamic tables list energy values for a standard state (25 °C =

298 K) and for different temperatures. These values are called Enthalpies
(J/mol) if the pressure is kept constant, normally at 1 atm.
To determine the energy necessary to produce aluminum two alternative
ways are possible (Lit. 5-1): the reaction takes place at the reference tem-
perature, followed by heating the products to the reacting temperature
(Figure 5-1)

5-4
5. Energy Balance
Figure 5-1: Reaction Path for Aluminum Production. The reaction takes place at room
temperature and the reaction products are heated to the reacting tempera-
ture.
or the reactants are first heated to the reaction temperature and subse-
quently allowing the reaction to occur at the new temperature (Figure 5-2).
Figure 5-2: Reaction Path for Aluminum Production. Contrary to Figure 5-1 the reactants
are first heated to the reaction temperature where then the reaction takes
place.
We follow the second alternative as Lit. 3-10 (p.19) is doing. The Elec-
trolysis Equation (Equation 2-13) is rewritten (divided by 4η) to produce
one mole of aluminum:
1
Al2 O 3 +
3 3 1
C = Al + 2 − CO 2 +
FG
3 1
− 1 CO
IJ FG IJ (5–3)
2 4η 4 η H
2 η K H K
The following enthalpy values have to be considered (see Chap. 18.12.2):
Enthalpies to heat from 25 °C to 970 °C:
Alumina: 108872 J/mol (Equ. 18-12-1),

Carbon Anodes: 17242 J/mol (Equ. 18-12-2).
Enthalpies of formation at 970 °C:
Alumina: 1690823 J/mol, (Equ. 18-12-3),

Carbon Dioxide: 395138 J/mol, (Equ. 18-12-4),
Carbon Monoxide: 113504 J/mol (Equ. 18-12-5).

5-5
5. Energy Balance
Using Equ. 18-12-7 the enthalpy of reaction at 970 °C is determined:
1 ⎛3 3 ⎞
ΔHreact = × ⎜ × 395138 − × 113504 ⎟ +
η ⎝4 2 ⎠
1 3 3
× 1690823 − × 395138 + × 113504
2 2 2
126098 (5-4)
ΔHreact = + 422961
η
and with Equ. 18-12-8 the total enthalpy to produce aluminum at 970 °C is:
126098 3
DH tot = + ¥ 17242 +
h 4h
1
422961 + ¥ 108872
2
139030 (5-5)
DHtot = + 477397
h
Converting the energy units from J/mol to kWh/kg Equ. 5-5 is rewritten to
yield the Enthalpy (Energy) required to produce One Kilogramm of Alu-
minum:
103 1
ΔHtot ( kWh/kg) = × 3 × ΔHtot ( J/mol )
26.9815 10 × 3600
1.431 (5–6)
ΔHtot = + 4.915
η
In this derivation the following effects were neglected:

Technical alumina consists mainly of gamma alumina. Gamma alumina is
transformed to alpha alumina when it comes into contact with the electro-
lyte. Since alpha alumina is more stable the heat of transformation is pro-
duced. However this effect is small, referring to Equ. 5-3 about 0.02
kWh/kg (see Lit. 3-10 Tab. 1-3):

5-6
5. Energy Balance
1000 2000
2 kJ / mol = × = 0.02 kWh / kg
27 1000 × 3600
In addition to the electrolytic part some additional carbon is consumed.

This anode excess consumption (AEC) is mostly due to airburn e.g. the
reaction of carbon anode with the ambient air. According to Lit. 5-1 the ex-
cess carbon is mostly combusted in the cell; it does not occur in the elec-
trolyte and will not contribute to the heat content of the bath.
According to the First Law of Thermodynamics the result of the considera-
tion is independent of the reaction path. The state of the reactants and of
the products is essential e.g. it does not matter how the chemical reaction
takes place in the electrolytic cell. Therefore the heat of dissolution of alu-
mina in the electrolyte or deviations from the ideal behavior must not be
considered.
5.4. Heat Loss

The difference between the added energy and the energy required to pro-
duce aluminum is lost as heat as already indicated in the introduction. Af-
ter defining the boundaries of the system (steel shell and alumina or
crushed bath cover) a steady state power balance is written considering
Equ. 2-16 for the amount and Equ. 5-5 for the energy of electrolytically
produced aluminum:
PInput = UHL ⋅ I = ΔH tot ⋅ PAl + Q

⎛ 1.431 ⎞
Q = UHL ⋅ I − ΔH tot ⋅ PAl = UHL ⋅ I − ⎜ + 4.915 ⎟ ⋅ 0.3356 ⋅ I ⋅ η ⋅ 1
⎝ η ⎠
Q = I ⋅ (UHL − 1.649 ⋅η − 0.480 ) (5–7)
For the pot voltage which contributes to the heat loss (UHL) normally the
pot voltage is taken reduced by the voltage drop of the external bus bars
and the anode voltage (see Chapter 4.2).
Assuming a pot voltage of 4.5 V and a current efficiency of 95 % shows
1648
. × 0.95 + 0.48
= 0.45
4. 5
that only 45 % of the electrical power is used to produce aluminum. The
rest is lost as heat. Considerable effort has been made and also great
progress has been achieved to reduce the heat loss of the cell (see Chap-

5-7
5. Energy Balance
ter 7). Optimisation of the thermal insulation of the bottom pot lining and
the side walls will save energy. The thermal layout of a pot should be de-
signed such that it maintains the correct bath temperature. The heat flow
should, however, be sufficient to maintain a layer of frozen electrolyte over
the side wall. This layer determines also the flow of the electric current e.g.
the magnetic effects like metal pad movement and upheaval (see Chapter
8). The thickness of this ledge depends on cell design and pot operation.
That is why great efforts are made to design a cell properly and to run it in
such a way that the optimal side ledge is formed.
Figure 5-3 shows a typical distribution of the heat loss of an electrolytic
cell. About half of the heat is lost on the upper part of the cell and the other
half on the side walls and the bottom.
Figure 5-3: Distribution of heat loss of a technical electrolytic cell.

5-8
5. Energy Balance
5.5. Reversible Decomposition Voltage

Thermodynamic uses besides the enthalpy another energy function the so
called Gibbs (Free) Energy. The following relation is written
ΔH = ΔG + T ⋅ ΔS (5–8)
Equation 5-7 states that the entire energy (ΔH enthalpy) of a reaction is
divided intowo parts: ΔG the minimum work and TΔS the amount of heat
which are necessary to execute a chemical reaction. Equation 5-7 is a
consequence of the Second Law of Thermodynamics. Several formula-
tions are found in the literature: it is impossible by any self-sustaining
process for heat to be transferred from a colder to hotter body or all natu-
ral processes will take place in such a way that entropy increases.
To show the meaning of Equation 5-7, let aluminum oxide be decomposed
chemically into aluminum and oxygen.
1 3
Al2O 3 = Al + O 2 (5–9)
2 4
For reaction (5–9) ΔH = 845 kJ/mol energy are necessary to produce one
mole of aluminum at 977 °C. Of this global energy ΔG = 640 kJ/mol must
be furnished as work e.g. electrical energy and the rest TΔS = 205 kJ/mol
as heat.
The Nernst Equation combines the voltage of an electrolytic process with
the Gibbs energy:
ΔG = ne ⋅ E 0 ⋅ F (5–10)
For reaction (5–9) one calculates the reversible decomposition voltage E0

knowing that three electrons are necessary:
ΔG 640 × 1000
E0 = = = 2.21V
ne ⋅ F 3 × 96485 (5–11)
E 0 = 2.21V

5-9
5. Energy Balance
This value E0 = 2.21 V applies for electrolytic processes which use inert
anodes (see Chapter 1-4).
Considering processes which produce energy for instance an electro-
chemical fuel cell which uses according to reaction (5–9) aluminum and
oxygen to produce electric energy the greatest useful work obtainable from
the process is equal to the Gibbs energy the rest of the energy is lost as
heat. The greatest voltage of such a battery would be 2.21 V .
For the case of the Hall-Héroult process the carbon anodes contribute to
reduce the reversible decomposition voltage. The Gibbs Energy for the re-
action
3
4
b
C + O 2 = CO 2 g (5–12)
is ΔG = 297 kJ/mol e.g. the reversible decomposition voltage E0 for the re-
action according to the Principal Equation (Equation 2-1)
2Al2O3 + 3C = 4Al + 3 CO2
is determined as
ΔG (640 − 297) × 1000
E0 = = = 118
. V
ne ⋅ F 3 × 96485 (5–13)
E 0 = 118
. V
The Hall-Héroult process reduces by using carbon anodes the minimum

work e.g. minimum voltage (reversible decomposition voltage) by about
one volt. The rest of the energy necessary for the production of aluminum
can by furnished as heat which is available in abundance.
The dependence of the reversible decomposition voltage of the Hall-
Héroult process on the temperature and the composition of the electrolyte
is given in Chapter 4.4.

5-10
5. Energy Balance
5.6. Literature
Lit. 5-1: J. J. J. Chen, M. P. Taylor, B. J. Welch,

"Thermochemical Energy Calculations for the Formation of Aluminium
by the Hall-Heroult-Process",
Erzmetall (1992) 45, Nr. 9, 468-470.

2nd Edition,
ISBN 3-87017-233-9.

5-11
5. Energy Balance

5-12
6.1. Summary ...................................................................... 6-3
6.2. Introduction ................................................................. 6-3
6.3. Mechanism of Metal Reoxidation............................... 6-4
6.4. Determination of Current Efficiency.......................... 6-6
6.5. Literature...................................................................... 6-8

6-1

6-2
6.1. Summary
In the course of the smelting process already produced aluminum is

lost. The reasons are discussed.
The reaction of aluminum with the anode gas is the primary reason
for current efficiency losses. The mechanism of this reoxidation re-
action is described.
Current efficiency is determined industrially by weighing the amount
of produced metal. To obtain current efficiency values for short time
periods the aluminum inventory is tracked by following the content
of tracer elements in the metal pad or by analyzing the composition
of the anode gas.
6.2. Introduction
In modern electrolysis cells the average current efficiency has reached
95 - 96 percent. The reasons for reduction of current efficiency in the
technical aluminum smelting process are (Lit. 6-1):
back reaction:
it is widely accepted that the reoxidation reaction (Equation 2-12) is the
primary reason for current efficiency losses.
depostion of more noble impurities from raw materials: iron for in-
stance as impurity is deposited electrolytically before aluminum. Other
elements like phosphorus or vanadium are discussed which are reduced
to a lower valency state at the cathode and oxydize afterwards at the an-
ode.
electronic conduction: at elevated temperatures solids and also liquids
may show electronic conduction. This electronic conduction acts like a
short circuit.
short circuiting: due to strong electromagnetic effects the oscillating
metal may touch the anodes provoking short circuits. Carbon pieces bro-
ken away from the anodes may cause physically short circuit.
leakage currents: the electric current may bleed through electric insula-
tion which is defective.
vaporization oxidation, absorption and reaction with container mate-
rials, physical losses: all reactions or manipulations which remove al-
ready produced aluminum metal contribute to the reduction of current effi-
ciency.

6-3
anode effects: during an anode effect the electrolysis mechanism is

changed. The bath is agitated and the temperature increased. This can
lead to considerable losses in current efficiency.
Practice has shown that the current efficiency is influenced by several pa-
rameters. For some influencing factors, rules of thumb are reported like
bath temperature: +0.2 % CE for -1 °C temperature decrease,
alumina content : +2 % CE for -1 % alumina content decrease.
For some other parameters the influence is accepted but no clear relation-
ships or diagrams are found in the literature: anode to cathode distance,
aluminum fluoride content, current density, height of metal pad. The rea-
son is probably that the influence of these parameters on the current effi-
ciency is indirect for instance when changing the thermal or electromag-
netic state of the electrolytic pot.
6.3. Mechanism of Metal Reoxidation

If aluminum is added to molten cryolite which is transparent like water im-
mediately fog-like streamers rise from the metal surface. Soon the melt
becomes opaque. The scholars are still in discussion about the type of
dissolution of aluminum in the electrolyte: reduced species, colloid or finely
divided aluminum. The assumption of colloidal dispersion of aluminum
seems the most reasonable.
Aluminum is transported away from the metal pad surface by convection
to the anode where it reacts with carbon dioxide. Figure 6-1 shows sche-
matically this reaction zone (Lit. 6-2 p.136) and Figure 6-2 the
Figure 6-1: Reaction zone between the liquid aluminum metal pad (cathode) and the
carbon anode.

6-4
corresponding concentration profiles of aluminum and carbon dioxide

(Lit. 6-3 p. 139).
Figure 6-2: Schematic concentration profiles for dissolved aluminum (c*Al, cAl, thermo-
dynamic and bulk concentration of aluminum) and dissolved carbon dioxide
(c*CO2, cCO2 thermodynamic and bulk concentration of carbon dioxide) in the
electrolyte.
The concentration of dissolved carbon dioxide is practically zero in the
bulk of the electrolyte. Aluminum must be transported by convection (or
diffusion) into the reaction zone near the anode. The concentration of alu-
minum in the bulk is therefore smaller than the thermodynamic saturation
value at the interface aluminum electrolyte.
Several models (some of them rather complex) for the reoxidation process
and for the current efficiency have been proposed (see the discussion in
Lit. 6-2 p. 139). The main parameters in the relationships are:
aluminum/electrolyte interface,
viscosity of the electrolyte,
relative velocity metal/electrolyte,
interelectrode distance,
density of the electrolyte,
aluminum saturation concentration,
electric current intensity.
These factors influence in essence the transport of dissolved aluminum to

the reaction zone. The model relationships do not contain however explic-
itly the electrolyte temperature which is thought to be one of the main fac-
tors determining the current efficiency. It seems that the temperature influ-
ences indirectly the current efficiency by changing the reaction parame-
ters.

6-5
6.4. Determination of Current Efficiency

Current efficiency is determined in industrial cells by weighing the amount
of metal tapped from the cell and comparing this with the integrated am-
pere-hours passed through the cell. Due to the large metal reservoir in the
cell, this method can only give the average current efficiency for periods of
several months.
To determine the current efficiency for short time periods two approaches
are used:
Dilution Technique
To analyse the aluminum inventory small amounts of an impurity are
added to the metal pad. Analysis of aluminum samples will give the alumi-
num mass in the cell. The requirements for a tracer material are: it must
form a homogeneous solution with the aluminum metal, it must not dis-
solve in the electrolyte or lining and it should not contaminate the outgoing
metal. Metals are used as tracer like copper, nickel, silver, gold but also
isotopes of copper, zirconium and gold were applied.
Accurate current efficiency determinations for short time intervals (weeks)
put strict requirements on the precision of the work especially on the
chemical or radioactive concentration measurements. To improve accu-
racy several tracers were added simultaneously in larger quantities to yield
higher concentrations which were tracked over the period of determination.
In this way one-week current efficiencies with one percent accuracy can
be obtained (see Chapter 14.4.2).
Gas Analysis
Gas analysis is used for instantaneous current efficiency measurements.
The common technique is to sample unburned, undiluted cell gas taken by
holes in the anodes and calculate the current efficiency using the Pearson-
Waddington relationship (Equation 2-39). Also samples of the exhaust
gas were used for the determination of current efficiency.
There are many possible errors in using the gas analysis technique:
aluminum can react with carbon monoxide to produce carbon,
2Al + 3CO = Al2O3 + 3C (6–1)
carbon dioxide can react with carbon dust in the electrolyte (Boudouard
Reaction),
CO2 + C = 2CO (6–2)
air leakage into the hot gas sample dilutes the sample and may oxidize
some carbon monoxide to carbon dioxide,

6-6
hydrocarbons from the anodes (especially in Söderberg pots) can produce

hydrogen and methane which alter the composition by the water-gas reac-
tion
H2 + CO2 = H2O + CO (6–3)
water which originates from moisture in the raw materials and from the air
reacts with aluminum fluoride of the electrolyte to add hydrogene fluoride
to the anode gas.
2AlF3 + 3H2O = Al2O3 + 6HF (6–4)
Several authors have modified the Pearson-Waddington equation taking

into account the side reactions. These modifications are discussed in
Lit. 6-4 (Chapter 9.5.1 p. 345-350) and Lit. 6-2 (Chapter 8.5 p.147-149).

6-7
6.5. Literature
2nd Edition,
ISBN 3-87017-233-9.
Lit. 6-2: A. R. Burkin (Editor),

"Production of Aluminum and Alumina",
Critical Reports on Applied Chemistry Volume 20,
John Wiley & Sons, (1987).
ISBN 0 471 91424 X.

"Understanding the Hall-Héroult Process for Production of Aluminum",
ISBN 3-37017-181-2.
Lit. 6-4: K. Grjotheim, C. Krohn, M. Malinovský, K. Matiašovský, J. Thonstad

"Aluminium Electrolysis, Fundamentals of the Hall-Héroult-Process",
2nd Edition (1982).
Aluminium Verlag, Düsseldorf;

6-8
7. Thermal Effects
7. Thermal Effects
7.1. Summary ...................................................................... 7-3
7.2. Introduction ................................................................. 7-3
7.3. Thermal Calculations .................................................. 7-5
7.4. Literature.................................................................... 7-15

7-1
7. Thermal Effects

7-2
7. Thermal Effects
7. Thermal Effects
7.1. Summary
The aim of the thermal layout of an electrolytic pot is to design the

thermal insulation e.g. the heat losses of a pot in such a way that an
optimal ledge is formed. No ridge under the anode should be devel-
oped and the ledge should protect the pot lining against the corro-
sive attack of the electrolytic bath. An optimal shape ledge limits the
flow of the electric current for minimal magnetic effects.
The solution of two coupled partial differential equations determines
the temperature distribution in an electrolytic cell. The interaction of
thermal effects with the magnetohydrodynamic behavior of the cell
strongly influence the results of the calculations. The comparison of
calculated and measureed results shows the validity of the method.
7.2. Introduction
The heat balance of a reduction cell is of great importance to achieve good
cell performance. Equation 7-1 shows the relationship between the pot
voltage and the specific energy consumption. To achieve low energy con-
sumption the voltage should be as low and the current efficiency as high
as possible:
U (7–1)
E = 297.97 ×
CE
the meaning of the symbols is:
The cell voltage can be easily reduced by moving down the anode beam
e.g. by reducing the anode to cathode distance (ACD). This can only be
done however in certain limits. If the anode to cathode distance (ACD) is
too low the metal pad - electrolyte interface becomes unstable due to
magnetohydrodynamic effects and metal reoxidation will occur. Hence,
current efficiency decreases and energy consumption increases according
to Equation 7-1. Therefore the values of the anode to cathode distance
should remain within a range at which the specific energy consumption is
a minimum.
Once the optimal value of the anode to cathode distance is determined the
amount of heat produced within the cell is determined. This heat is pro-
duced by the Joule effect which depends on the anode to cathode dis-
tance but also on the electrical resistances of the anodes, the carbon lin-

7-3
7. Thermal Effects
ing, the collector bars and on contacts. However, about two thirds of the
heat is produced in the anode to cathode distance.
The thermal insulation of the cell must be designed accordingly so as to
dissipate properly this amount of heat. Indeed if the cell is not sufficiently
thermally insulated a ledge will grow on the side lining and will eventually
reach under the anodes leading to a strong ridge. Figure 7-1A shows a
standard ledge on the pot side. Figure 7-1B in contrary shows a strong
ledge formation. The presence of ridge is also directly correlated to a de-
crease of current efficiency. On the other hand, if the energy input into the
cell is too big or if the cell is to much thermally insulated, ledge will melt
away especially on the side lining (Figure 7-1C). The carbon side blocks
will be mechanically and chemically attacked especially at the electrolyte -
metal pad interface.
(A) (B) (C)
Figure 7-1: Different types of ledge formation: (A) standard ledge, (B) strong ledge forma-
tion with ridge under the anode, (C) missing ledge on the potside.
Aim of the thermal layout of an electrolytic pot is to design the thermal in-
sulation e.g. the heat losses of a pot in such a way that an optimal ledge is
formed. The ledge should protect the pot lining and limit the flow of the
electric current for minimal magnetic effects.
In Figure 7-2 steps are shown which are necessary to design the thermal
pot isolation:

7-4
7. Thermal Effects
Figure 7-2: Steps necessary to design the thermal insulation of an electrolytic pot.
7.3. Thermal Calculations
The determination of the temperature distribution in the cell, ledge profile

and heat losses is realized by solving the coupled electrical partial
differential equation (Equation 7-2):
div ( σ grad ϕ ) = 0 (7–2)
and the thermal partial differential equation (Equation 7-3):
∂T
rC p
∂t
b g 2
= div lgrad T - rCp V grad T + s gradj +
df s (7–3)
+ rDHPhase + Dh Chem
dt
the meaning of the symbols in Equation 7-2 and 7-3 is:

7-5
7. Thermal Effects
Discussion of the Electrical Partial Differential Equation (Equa-

tion 7-2):
The local current density can be expressed in terms of the electrical poten-
tial:
a f
J x , y , z = −σ grad ϕ ≡ −σ
FG ∂ϕ , ∂ϕ , ∂ϕ IJ (7–4)
H ∂x ∂y ∂z K
Since the number of electrons in any small volume element is constant

e.g. there is no accumulation of charge in the electrolysis cell Equation 7-5
is valid and equivalent to Equation 7-2:
∂J x ∂J y ∂J z
a
div σ grad ϕ = div J ≡ f af ∂x
+
∂y
+
∂z
=0 (7–5)
Equation 7-4 is tightly related to the well known Ohm’s law. Consider for
example a conductor of cylindrical shape between two blocs of highly con-
ductive material:
Conductivity: σ
ϕin Section: S ϕout
J
Length: L
In terms of the total current I=SJ, Equation 7-4 can be written as:
ϕ in − ϕ out L
= ≡R
I σS (Ohm’s law)

7-6
7. Thermal Effects
where R is the resistance of the cylindrical conductor.
Discussion of the Thermal Partial Differential Equation (Equation 7-3):

Right Hand Side of Equation 7-3:
∂T
ρC p
∂t
The right hand side of Equation 7-3 gives the variation of the heat density
in time at the location (x, y, z) in the cell. To this variation corresponds a
temperature variation, which appears in the left hand side of Equation 7-3.
The contributions to the variation of heat density on the right-hand side of
Equation 7-3 are:
Diffusive Heat Transfer:

a
div λ gradT f
this term describes the variation in time of the heat density at the location
(x, y, z) due to diffusive heat transfer. More specifically, the density of the
heat flow rate due to diffusion is given by
a f
q x , y , z = − λ gradT (7–6)
Convective Heat Transfer:
− ρ ⋅ CpV gradT ≡ − ρ ⋅ Cp Vx
∂T FG
+ Vy
∂T
+ Vz
∂T IJ
∂x H ∂y ∂z K
this term describes the variation of the heat density at location (x, y, z) due
to the convective heat transfer e.g. the heat transported by fluid flow. The
velocity field V(x, y, z) is non-zero only in regions occupied by the electro-
lyte and the metal pad.

7-7
7. Thermal Effects
Heat Generated by Joule Effect :

2
s ◊ grad j ∫ s -1J 2
describes the heat generated by Joule effect (i.e. due to the electrical cur-
rent and the ohmic resistance). This is a heat source term. The ratio of
electrical conductivity in the liquid metal to the electrical conductivity in the
bath is close to 16000 e.g. the conductivity σ of bath is much lower than
any other conductive material in the cell. Most of the heat is produced in
the anode to cathode distance.
Heat Generated by Phase Transformation:

dfs
ρΔH Phase
dt
describes the heat generated by the freezing of the bath near the ledge
due to the latent heat of solidification. This is a heat source term.
Heat Generated by Chemical Reactions:

ΔhChem
describes the heat generated by chemical reaction (e.g. the Boudouard
reaction on the faces of the anode). This is a heat source term which plays
a minor role.
Note that the thermal Equation 7-3 is analogous to the electrical Equation
7-2 in the case of stationary heat flows and when thermal sources and
convective heat transports are absent. Indeed, in such conditions, the
temperature T, the heat flow q and the thermal conductivity λ play the
same roles as the electrical potential ϕ, the current density J and the elec-
trical conductivity σ and it is thus possible to write a thermal Ohm’s law in-
volving thermal resistances. Consider for example a solid thermal conduc-
tor of cylindrical shape between two blocs with high thermal conductivities:
Conductivity: λ
Tin Section: S Tout
q
Length: L
In terms of total heat flux Q=Sq, Equation 7-6 can be written as:
Tin − Tout L
= ≡ Rth
Q λS (Thermal Ohm’s law)
where Rth is the thermal resistance of the cylindrical conductor (see Exer-
cise 14-1).

7-8
7. Thermal Effects
Unknown values for Equation 7-2 and 7-3 are the electrical potential ϕ and
the temperature T. In order to solve the problem, the equations must be
completed with the boundary conditions, that is to say that the current
flowing in each anode must be specified and the potential at the end of
each collector bar must be imposed,
One set of possible boundary conditions imposed on the heat flow is
specified by:
a fe a
q b = hb x , y , z ,T Tb x , y , z − Ta f j (7–7)
Equation 7-7 states that the heat flow at the boundaries is proportional to
the temperature difference between the cell surface and the pot room with
the restriction, however, that the heat exchange coefficient hb(x,y,z,T) de-
pends on the temperature. Once the boundary conditions are known
Equation 7-2 and 7-3 can be solved using, for example, the following itera-
tive procedure:
Figure 7-3: Iterative procedure to solve Equation 7-2 and 7-3.

If the material properties are defined, the problem can be solved using the
Finite Element Method. The problem can be solved either to find the cell
stationary state or to study non stationary states such as cell startup or
any other time changes.

7-9
7. Thermal Effects
At stationary state the heat generation in the cell should be equal to the
heat losses. The calculation of both allows to control the accuracy of the
numerical simulations. The problem is highly non linear. The non linearity
is caused by :
• Thermal boundary conditions (not critical)
• Electrical conductivity is to a high degree non linear in metal and bath.

Indeed, when the temperature is lower than about 940 ºC both metal
and bath are transformed in so called ledge which is an insulating mate-
rial. Between both phases, a mushy zone is assumed in which the elec-
trical conductivity is supposed to change linearly to a value close to
zero.
• Thermal conductivity is to a high degree non linear in metal and bath.

Indeed, when the temperature is lower than about 940 ºC both metal
and bath are transformed in so called ledge which is a material with
much lower thermal conductivity. Between both phases, a mushy zone
is assumed in which the thermal conductivity is supposed to change
linearly to a value of 1.16 W/mK.
• Electrical and thermal conductivity of other materials depend only

slightly on temperature and most of them do not depend significantly on
temperature.
A typical example of thermal calculation results is shown in Figure 7-4.
Figure 7-4: Isotherm in a three dimensional representation of a cell.
The results can further be analyzed in sections of the cell (Figure 7-5).

7-10
7. Thermal Effects
(A) (B)
Figure 7-5: Ledge formation in vertical slices of one quarter of the cell. (A) in the trans-
verse and (B) in the longitudinal direction.
Another difficulty for solving the problem comes from the convection term
which needs the knowledge of the velocity field in both the metal and elec-
trolyte. The velocity field is a result of the magnetohydrodynamic (MHD)
calculations (Lit. 7-1, 7-2), which themselves depend on the thermal calcu-
lations. Modifications to the ledge change the current density in the metal
pad which in turn affects the velocity field. Figure 7-6 and 7-7 show the ve-
locity field in a horizontal plane at the middle of the metal pad for two dif-
ferent ledge shapes. The impact of the presence of ledge under the an-
odes (ridge) is obvious.
Figure 7-6: Velocity field in a horizontal plane at the middle of the metal pad assuming no
ridge. The maximal speed of metal pad motion was Vmax = 21 cm/s.
Figure 7-7: Velocity field in a horizontal plane at the middle of the metal pad assuming a
pronounced ridge. The maximal speed of metal pad motion was
Vmax = 33 cm/s.

7-11
7. Thermal Effects
There is therefore a coupling between the magnetohydrodynamic calcula-

tions and the thermal calculations.
Figure 7-8: Visualization of existing interactions between the magnetohydrodynamic ef-

fects and the thermal balance of the cell.
Figure 7-8 contains three circles: the outer circle with thermal insulation,
magnetic fields etc. refers to the design and conditions of pot operation.
The intermediate circle with heat loss, ledge shape etc. refers to the dy-
namic variables of the cell. The inner circle with pot age, current efficiency
etc. refers to the cell performance.
Figure 7-8 shows how cell design and pot operation (outer circle) influence
the dynamics of the cell (intermediate circle) which in turn determine the
performance of the cell (inner circle). The anode to cathode distance de-
termines the cell voltage which for a given current determines the heat
loss of the cell. The heat loss together with the design of the thermal insu-
lation of the cell determine the ledge shape which in turn influences the
current distribution in the cell. In the presence of a magnetic field (gener-
ated by electrical currents in the external bus bars and in the cell) this cur-
rent distribution induces fluid motions whose velocity depends strongly on
the current distribution itself. The presence of a ridge, for instance, under
the anode leads to important horizontal currents which results in fluid ve-
locities destabilize the metal pad (noise). Finally the presence of noise
would force the operator to increase the anode to cathode distance.
In order to avoid to repeat the whole magnetohydrodynamic calculations
at each step of the thermal calculations, it is interesting to analyze the in-
fluence of the convection term in the thermal equation or in other words to
analyze the Peclet Number. A discussion about this can be found in
Lit. 7-3. The Peclet Number (Pe) is the ratio of the amount of heat trans-
ported by bulk convection to the amount of heat transported by diffusion:

7-12
7. Thermal Effects
bulk heat transfer ρC pVdl (7–8)

Pe = =
conductive heat transfer λ
The following typical values for the electrolyte:
lead to a rather high Peclet number:
2100 × 1650 × 0.05 × 0.01

Pe = ≈ 43000
0.4
The value of Pe = 43000 indicates that the heat is mainly transported by

convection.
The influence of the velocity field has been analyzed in Lit. 7-4. It was
shown that good results can be obtained when increasing artificially the
thermal conductivity instead of taking the velocity field into account. How-
ever, local effects due to the velocity field can only be studied by analyzing
the global magnetohydrodynamic and thermal problem. It appears that the
ledge shape is much more sensitive to the velocity field in the bath than in
the metal. However, it cannot be neglected for either fluid.
Finally one should mention that there are many parameters which play an
important role on thermal equilibrium. The effects of metal height, number
of cathode bars, external temperature, thermal insulation, etc. Some as-
pects have been discussed in Lit. 7-3.
The influence of the current load on the ledge shape is presented as an
example in Figure 7-9 for constant anode to cathode distance, anode di-
mensions and thermal insulation.
Increasing the current in a potroom seems to be an easy and cheap way
to improve the productivity of a plant, assuming that there will be no sur-
prising effects on current efficiency, energy consumption and cell life. The
stability of the process depends on a good pot control system if the cell is
not in thermal equilibrium. The ledge plays an important role because it
must protect the lining for improving the cell life, but it must not grow under
the anodes. Indeed, if a ridge exists, the cell voltage is increased due to an
increase of the cathode voltage drop and to an increased noise level. It is
therefore very important to be able to predict the ledge shape. Figure 7-9
A to D show the influence of the current intensity on the ledge profile and
the bath temperature keeping all the other parameters constant. The cur-

7-13
7. Thermal Effects
rent density increases from 0.70 A/cm2 to 0.87 A/cm2. The ledge melts
partially in order to satisfy the thermal balance.
(A) (B)
(C) (D)
Figure 7-9: Influence of the current load on the bath temperature and the ledge shape.
(A) current intensity I = 120 kA, bath temperature TB = 958 °C, (B) I = 125 kA,
TB = 962 °C, (C) I = 130 kA, TB = 965 °C, (D) I = 135 kA, TB = 971 °C.
Figure 7-10 compares a measured ledge profile, averaged for five cells
with the corresponding calculated ledge. The measurements have been
performed with a ledge scanner (Lit. 7-5). The agreement is reasonably
good. It should be mentioned, however, that shell surface temperature
measurements were performed in order to carefully check the heat ex-
change coefficient prior to calculating the ledge shape.

7-14
7. Thermal Effects
Figure 7-10: Comparison of the average ledge shape determined on five production cells
with the calculated ledge.
Three-dimensional thermal calculations of the Hall-Heroult cell can be

used for predicting the temperature distribution, ledge profile, heat losses,
and many other cell behavior. Lit. 7-6 to 7-8 are recently published papers
which may be of interest in this context because they deal with the thermal
behavior of Hall-Heroult cells.
7.4. Literature
Lit. 7-1: J. Antille, Y. Kraehenbuehl, R. von Kaenel, J. C. Weber,
"Fluid Flow Control: A Must for the Aluminum Industry"
Light Metals 1992, 1247-1256
Lit. 7-2: J. Antille, R. von Kaenel, M. V. Romerio,

"On the Effect of Induced Currents in Electrolysis Reduction Cells",
Proceedings of the International Symposium on Advances in Production and
Fabrication of Light Metals and Metal Matrix Composites
Edmonton, Alberta, August 23-27 1992, p.63-73
Lit. 7-3: J. N. Bruggeman, D. J. Danka,

"Two-Dimensional Thermal Modeling of the Hall-Heroult Cell",
Light Metals 1990, 203-209
Lit. 7-4: J. Antille, M. Givord, Y. Kraehenbuehl, R. von Kaenel,

"Effects of Current Increase on Aluminium Reduction Cells",
Lit. 7-5: W. Schmidt-Hatting, J.-M. Blanc, J.-C. Bessard, R. von Kaenel,

"Heat Loss of Different Pots",
C. C. Wei, J. J. J. Chen, B. J. Welch,

"Modelling of Dynamic Ledge Heat Transfer",
Lit. 7-6: Light Metals 1997, 309-316
Lit. 7-7: W. Haupin,

"Understanding Boundary Layers",
Lit. 7-8: A. Solheim, L. Støen,

"On the Composition of Solid Deposits Frozen out from Cryolite Melts",

7-15
7. Thermal Effects

7-16
8. Magnetic Effects
8. Magnetic Effects
8.1. Summary ...................................................................... 8-3
8.2. Introduction ................................................................. 8-3
8.3. Stationary state ........................................................... 8-4
8.3.1. Introduction ........................................................................... 8-4
8.3.2. Application .......................................................................... 8-10
8.4. Stability Analysis....................................................... 8-12
8.4.1. Introduction ......................................................................... 8-12
8.4.2. Theoretical Approach .......................................................... 8-15
8.4.3. Applications......................................................................... 8-21
8.5. Literature.................................................................... 8-27

8-1
8. Magnetic Effects

8-2
8. Magnetic Effects
8. Magnetic Effects
8.1. Summary
• The Hall-Héroult electrolysis process is very complex. It involves

many different physical and chemical phenomena interacting with
each other. In general, we can fit Hall-Héroult mathematical models
into three categories:
• thermal-electrical models associated with the cell heat balance
(see Chapter 7),
• magnetohydrodynamic (MHD) models associated to the calcula-
tion of the ”magnetic state” of the cell,
• stress models generally associated with cell shell and cathode de-
formation.
• Those aspects of cell design interact in different ways:
• MHD is affected by the ledge profile, which is defined by the cell
heat balance,
• local ledge profile is affected by the metal circulation pattern
mostly dictated by the busbar design.
• The goal of the magnetic modelling is to optimize the design of the
busbar system and operation of aluminum reduction cells, in order
to achieve higher current efficiency, cell life and lower specific en-
ergy consumption. Magnetic modelling is not only the determina-
tion of the magnetic field, it also includes the current flow deter-
mination, the liquid metal and bath velocities and the deformation
of the metal pad surface. Magnetohydrodynamic instabilities or in
other words “noise level”, is a dynamic model to compute the an-
swer of the cell due to process perturbations (current increase,
anode change, metal tapping, busbars…). It is extremely important
to validate these models and a lot of time has been spent to com-
pare the calculated with the measured data from both a stationary
and an unstationary point of view.
8.2. Introduction
The understanding of electromagnetic interactions in Hall-Héroult cells, in
particular the stability of the metal surface in alumina reduction cells has
been one of the major factors for the improvements of cell performance,
such as current efficiency, energy consumption and life of the reduction
cells.
The use of modern modelling tools allowed to optimize both thermal and
magnetohydrodynamic states of the cells. Optimization of the busbar sys-
tem is achieved by means of computer simulations which take into account

8-3
8. Magnetic Effects
the local arrangement of busbars, their current density as well as the ar-
rangement of all relevant steel parts. Based on computer simulation, the
stability analysis allows to determine the maximum current load of a given
line.
8.3. Stationary state
8.3.1. Introduction
Many papers can be found in the literature on the effects of magnetic

fields in alumina reduction cells (Lit. 8-1 to 8-10). The present chapter is
an attempt to discuss several aspects of the influence of magnetic fields
but it is not the intention to describe in detail the mathematical equations of
the modelling.
Figure 8-1 summarizes the methodology associated to the calculation. The
first step consists in looking at the thermal equilibrium of the cell. The
ledge shape is therefore taken into account for the generation of a three
dimensional mesh. The mesh program contains geometric quantities as-
sociated to the real cell, such as the true shape ledge profile, the anodes,
fluids, cathodic blocs and the cathodic busbars.
Figure 8-1: Methodology associated to the magnetohydrodynamic calculations.

8-4
8. Magnetic Effects
The Poisson equation for the voltage drop is solved :

(8–1)
div(σ grad ϕ ) = 0
One can apply different boundary conditions for Equation 8-1, such as a
fixed potential at the top of the anode rod and at the end of the outgoing
cathode bars or a current distribution in the collector bars and anodes.
The current distribution inside the cell is calculated applying Ohm’s law
which is written in the vectoriell way:
(8–2)
j = σ ( − grad ϕ + V × B )
The induced current due to the fluid velocity field is taken into account.
The current is conserved within the cell and the current density is a vector
pointing vertically downward in the bath, but may have considerable spa-
tial non uniformity due to the cell design and the cell operation. The hori-
zontal component of the current, which has a great impact on the stability
of the cell, is influenced mainly by the frozen side ledge profile and its
variations.
The magnetic field in the bath and in the metal regions generated by the
electrical currents circulating along the external busbars is calculated by
Ampere’s law. This part of the magnetic field H is called the external field
He.
He = 1
z
j (r ′ ) × (r − r ′ )
4π Ω (r − r ′ ) 3
dΩ(r ′ ) (8–3)
Where Ω is the calculation domain.

In order to obtain the magnetic field inside the liquid regions generated by
the electric current distribution within the cell, a volumetric integration of
the current density is performed. This part of the magnetic field H0 is called
the internal field Hi.
Finally the total magnetic field H0 is calculated by adding both contribu-
tions:
H0 = Hi + He
(8–4)
For the determination of the induction field B, the influence of the steel
parts is taken into account. The induction field pattern is distorted by the
ferromagnetic elements like the steel shell, collector bars and yokes.
These steel parts are magnetized if their temperature is below the Curie

8-5
8. Magnetic Effects
temperature (770 °C). The nonlinear properties of ferromagnetic materials

are illustrated in Figure 8-2, showing how B is related to H for typical steel
used for the shell.
Figure 8-2: Curve of primary magnetization for the steel shell.
B is related to H by the following formula:
B = μr ⋅ μ 0 ⋅ H (8–5)
The solution of this nonlinear problem is achieved by an iterative process.

The permeability is adjusted to the magnetic field amplitude for each finite
element. As an example, Figure 8-3 shows the relative permeability within
the steel shell.

8-6
8. Magnetic Effects
Figure 8-3: Relative magnetic permeability of the steel shell.
Figure 8-4 and Figure 8-5 compare the three components of the magnetic
field Bx, By, Bz determined by calculation and measurement. The calcula-
tions were executed with and without considering the influence of the steel
parts of the cell.
Measurements were made in a potroom with pots arranged end to end.
The magnetic fields were measured on two pots at their long sides (Figure
8-4) and short sides (Figure 8-5). As you can see from both figures the of
calculated and measured results fit quite well.

8-7
8. Magnetic Effects
Figure 8-4: Comparison of measured and calculated values of the three components Bx,
By, Bz of the magnetic field for an end to end arrangement of pots. The calcu-
lations were executed with and without consideration of steel parts. The
measurements were taken for two pots on their long sides in the middle and
on the sides of the potroom. I indicates the direction of the line current.

8-8
8. Magnetic Effects
Figure 8-5: Similar to Figure 8-4 the measured and calculated values of the three compo-
nents Bx, By, Bz of the magnetic field for an end to end arrangement of pots
are compared. The calculations were executed with and without consideration
of steel parts. The measurements were taken for two pots on their short sides
up- and downstream corresponding to the direction of the current (I).
The electromagnetic force per unit volume is given by the vector product
of the current density and the induction field:
F = j ×B
(8–6)

8-9
8. Magnetic Effects
The electromagnetic force generates the motion in the liquid metal and
bath and deforms the metal-bath interface. The fluid motion is described
by the Navier-Stokes Equation:
∂u
ρ⋅ + ρ (u , ∇)u + ∇ ( p + ρgz ) = ηΔu + F (8–7)
∂t
The Navier-Stokes equation must be solved simultaneously in the bath

and in the metal for the velocities and the pressure. At the beginning, the
metal-bath interface is assumed to be flat. At each iteration, all above
mentioned equations are solved and the interface is deformed in order to
obtain a true equilibrium of the pressure through the interface. Normally
five iterations are necessary to obtain a good balanced pressure.
8.3.2. Application
The busbar design influences the circulation pattern in the cells. An exam-
ple of computed metal velocities for end-to-end and side to side pots is
shown in Figure 8-6 and Figure 8-7 respectively.
Figure 8-6: Circulation pattern of the metal pad determined for an end to end arrange-
ment of electrolytic pots.

8-10
8. Magnetic Effects
Figure 8-7: Circulation pattern of the metal pad determined for a side to side arrangement
of electrolytic pots.
For the development of reduction cells with a current load of more than
170 kA, the side-by-side configuration is used almost exclusively. The rea-
sons for this are mainly cost-saving and lower magnetic field due to
greater flexibility by manipulating the busbar layout.
The metal level varies from place to place over the metal surface due to
magnetic effects. These effects depend among other things upon busbar
and riser arrangement, pot age, anode current, metal volume and side
crust formation. Figure 8-8 shows the calculated and measured metal-bath
interface along the long side of a pot assuming that the current distribution
of the collector bars is perfect.
Figure 8-8: Metal pad interface contour. The interface was calculated and measured
along the long side for end to end configuration.
The shape of the metal surface depends on the collector bar current distri-
bution. Figure 8-9 shows the impact due to a poor collector bar current dis-

8-11
8. Magnetic Effects
tribution corresponding to a poor cell state. The numerical simulation of the

metal-bath interface shape is accurate enough to predict an optimum an-
ode setting.
Figure 8-9: Metal upheaval in the middle for a poor and ideal current distribution in the
collector bars.
8.4. Stability Analysis
8.4.1. Introduction
Stability of the electrolysis process is an essential factor for achieving high

current efficiency and low energy consumption. It is well known that the
surface of the molten aluminum in Hall-Héroult cells is not totally quiet. In
other words one can observe voltage fluctuations usually called “noise”.
The oscillation of the metal pad can be recognized by observing the cur-
rent fluctuation for each individual anode. As an example Figure 8-10
shows measurements of the anode current fluctuations for side by side
cells operated at 185 kA.
One can observe periodic oscillations in the range of 40 seconds which
are a typical indication of magnetohydrodynamic instabilities, also called
metal pad roll.

8-12
8. Magnetic Effects
Figure 8-10: Fluctuation of anodic currents with time. The individual current intensities of
the anodes were measured on a pot arranged side by side and operated at
185 kA.
The Fourier analysis of these measurements is performed in order to ob-
tain the frequency spectrum (Figure 8-11).

8-13
8. Magnetic Effects
Figure 8-11: Frequency spectrum corresponding to the anode current of Figure 8-10.
This technique exhibits the eigen frequencies of the oscillations of the

metal-bath interface.
The stability of the metal-bath interface has been studied since the early
seventies and the impact on cell operation has been reported in many pa-
pers (Lit. 8-11, 8-12, 8-13). Let us mention very simple results when as-
suming no current is flowing in the cell and there is a rectangular geome-
try. This approach allows to introduce the concept of modes associated
with eigen frequencies (Figure 8-12).

8-14
8. Magnetic Effects
Gravitational waves are defined in the following way (Figure 8-12):
2
ωm
a
g ⋅ ρ1 − ρ 2 ⋅ k f (8–8)
,n =
a f
ρ1 ⋅ cth k ⋅ h1 + ρ 2 ⋅ cth k ⋅ h2a f
2 2 2 m ⋅π n ⋅π
k = kx + ky , kx = , ky =
Lx Ly
Figure 8-12: Rectangular cell geometry corresponding to Equation 8-8 to treat gravita-
tional waves.
This rectangular approach (Lit. 8-9) for the geometry of the cell (three lay-
ers are considered: liquid metal, bath and anodes), does not take into ac-
count the true geometry of the cell. It is known from practice that the ledge
shape plays an important role on the stability. If the ledge grows under the
anodes (ridge), strong horizontal currents are generated and the cell be-
comes unstable.
8.4.2. Theoretical Approach
This section studies the stability of the linearized magnetohydrodynamic

(MHD) equations which govern - in the neighborhood of a stationary solu-
tion - the motion of two fluids in an electrolysis cell for the production of
aluminum. Our approach is based on the following experimental observa-
tion: the Fourier analysis of time series recording the anode currents of an

8-15
8. Magnetic Effects
oscillating cell shows that the frequencies correspond to the dominant

modes and are close to some of the fundamental frequencies which one
can compute for the cell under the action of gravity only. The calculations
scheme is as follows:
• the evolutionary MHD equations,

• the MHD system is then linearized in the neighbourhood of a stationary
solution,
• an iterative scheme for the computation of the mode fields and of the
frequencies.
We introduce the following notations for the evolution variables:
u~ ( x , t ) , velocity field,
~
p ( x, t )
, pressure field,
~
b ( x, t ) , induction field,
~j ( x , t )
, electric current density,
j~(x ,t ) , electric potential field,
~
h ( x, y, t ) , function describing the interface.
The evolution of the system is described by the following equations
1 ~ ~
ρ∂ t u~ + ρ (u~ , ∇)u~ + ∇( ~p + ρgz ) = rotb × b (8–9)
μ0
divu~ = 0
(8–10)
~
rotb = μ 0 ~j (8–11)
~ + u~ × b~ )
~j = σ ( −∇ϕ
(8–12)
div ~j = 0 (8–13)
~
divb = 0 (8–14)
{~
p}Γ ( h~ ) = 0 on ~
Γ(h ) (8–15)
~ ~ ~ ~
∂ t h − (u~ , ∇( z − h )) = 0 on Γ(h ) (8–16)
In Equation 8-9 to 8-16 the standard notations was used:

8-16
8. Magnetic Effects
Equation 8-16 expresses the immiscibility of the two fluids.

Now, we assume that a stationary solution exists and is given by the fields
u( x ), p( x ), h( x , y ), j ( x ), b( x ), ϕ ( x ) (8–17)
We consider time dependent solutions of Equation 8-9 to 8-16 which are

close to the stationary solution of 8-17. We assume that the evolution
fields are the sum of two fields :
u~ ( x , t ) = u( x ) + U ( x , t )
~
p ( x , t ) = p( x ) + P ( x , t )
~ (8–18)
h ( x , y , t ) = h( x , y ) + H ( x , y , t )
~j ( x , t ) = j ( x ) + J ( x , t )
~
b ( x , t ) = b( x ) + B ( x , t )
ϕ~ ( x , t ) = ϕ ( x ) + Φ( x , t )
The fluctuations U, P, H, J, B, Φ are supposed to be sufficiently small for

being submitted to a linearization procedure. We consider time dependant
solutions of the form
U ( x , t ) = eiωt U ( x )
P( x , t ) = eiωt P( x )
H ( x , y , t ) = eiωt H ( x , y ) (8–19)
J ( x , t ) = eiωt J ( x )
B( x , t ) = eiωt B( x )
Φ( x , t ) = eiωt Φ( x )
Substituting the Equation 8-18 and 8-19 in Equation 8-9 to 8-16 amounts
replacing ∂tU by iωU and ∂tH by iωH, we obtain the following system for the
fluctuation fields:
1
iωρU + ρ (u, ∇)U + ρ (U , ∇)u + ∇P = (rotb × B + rot B × b) (8–20)
μ0
divU = 0
(8–21)
{ P + H∂ z p}Γ = 0 on Γ (8–22)
iωH − (∂ z u, ∇( z − h)) H − (U , ∇( z − h)) + (u, ∇H ) = 0 on Γ (8–23)

8-17
8. Magnetic Effects
z
Ω
ρ
1
(∇ P , ∇P ) = 1 (8–24)
rot B = μ 0 J D
(8–25)
where J D = J + αH { j}Γ δ Γ , α = (1 + (∂ x h) 2 + (∂ y h) 2 ) −1/ 2 (8–26)
J = σ ( −∇Φ + U × b + u × B)
(8–27)
div J D = 0
(8–28)
div B = 0
(8–29)
{Φ + H∂ zϕ}Γ = 0 on ~
Γ (h ) (8–30)
To solve Equation 8-20 to 8-30, we introduce an iterative scheme for

computing eigenvalue ω and the corresponding mode fields U, P, H, J,
B, Φ.
We describe the (n+1) step. ωn and the fields Un, Pn,... are known and we
want to compute the fields:
ω n +1 = ω n + δω , U n +1 = U n + δU , P n +1 = P n + δP,... (8–31)
Writing Equations 8-14 to 8-23 with Un+1 = Un + δU, Pn+1 = Pn + δP,... we

then neglect some of the quadratic terms in δU, δP,... we obtain for δω, δU,
δP, δH the following linear equation system:
iρ (ω n U n + δω U n + ω nδ U ) + ρ (u, ∇)U n + ρ (U n , ∇)u + ∇( P n + δP) = (8–32)

1 n n
(rotb × B + rot B × b)
μ0
div (U + δ U ) = 0
n
(8–33)
{ P n + δ P + ( H n + δ H )∂ z p} Γ = 0 on Γ (8–34)
i(ωn Hn +δωHn +ωnδH) − (∂ z u, ∇(z − h))Hn − (U +δU, ∇(z − h)) + (u, ∇Hn ) = 0
n
(8–35)
on Γ

8-18
8. Magnetic Effects
z
Ω
ρ
1
{(∇ P n , ∇P n ) + 2(∇P n , ∇δP)} = 1 (8–36)
rot B = μ 0 J
n n
D (8–37)
where J D n = J n + αH n { j}Γ δ Γ , α = (1 + (∂ x h) 2 + (∂ y h) 2 ) −1/ 2 (8–38)
J = σ ( −∇Φ n + U × b + u × B n )
n n
(8–39)
div J D = 0
(8–40)
n
div B = 0 (8–41)
~
{Φ n + H n∂ zϕ}Γ = 0 on Γ (h ) (8–42)
The calculation of the initial approximation ω0, U0, P0, H0 is discussed in

detail in (Lit. 8-15).
In order to consider the real ledge shape in the stability analysis, the calcu-
lations of the metal pad oscillations have been coupled to the magnetohy-
drodynamic and thermal calculations (Lit. 8-14). The geometry of the pool
containing metal and bath is no longer rectangular. This fact leads to over-
come severe mathematical difficulties. The mathematical developments
are published in (Lit. 8-15, 8-16, 8-17, 8-18). The impact of the real ge-
ometry of a cell is presented in Table 8-1.
Table 8-1: Example of Eigenfrequencies for Gravitational Oscillations.
The left hand side periods are calculated by assuming gravitational waves
in a rectangular box, the right hand side periods are determined numeri-
cally on the basis of the real cell geometry.

8-19
8. Magnetic Effects
Let us summarize the basic assumptions and results:
Assumptions
• There exists a stationary state whose variables remain constant with

time.
• When fluctuations exist around a given stationary state, they are suffi-
ciently small to allow a treatment through linearization procedures.
• The fluctuations arising around the considered stationary state can be
described without taking into account the Stoke's forces (absence of
viscosity for fluctuations only).
• There is no constraint on the cell geometry.
The absence of viscous terms in the equations for the fluctuations is justi-
fied by experimental observations.
The free geometry leads to important mathematical complications. The
theory of distributions and a perturbation method are used. On the other
hand it should be mentioned that all the stationary state parameters like
current density, magnetic fields and velocity fields and the interface defor-
mation are taken into account for the stability analysis.
The stability of the cell for a given mode is given by the imaginary part of
ω. If it is positive or zero the cell is said to be stable, if it is negative, the
cell is said unstable. The negative values of the imaginary part are also
called growth factors as they define how fast a perturbation grows. Figure
8-13 shows the principle of a stability diagram.
Figure 8-13: Principle of a stability diagram .
This theory is a generalization of an algorithm published earlier. As we

see, the stability does not depend only on the magnetohydrodynamic
(MHD) state of the cell but also on its thermal equilibrium. Figure 8-14 (see
also Fig. 7-8) represents the fundamental interactions between MHD and
thermal equilibrium.

8-20
8. Magnetic Effects
Figure 8-14: Interactions between magnetohydrodynamic effects and thermal equilibrium

of the cell.
When the energy input is modified by changing the voltage or the current,
the heat losses are modified and, as a consequence, the ledge shape
changes. The ledge shape itself varies with the current density, in particu-
lar at the metal level, which implies a change of the induction of the mag-
netic field. The magnetic field induces the velocity field and the metal up-
heaval. According to theory, all these quantities define the cell stability. Fi-
nally the operational parameters allow to modify the cell state by changing
the ACD or the voltage set value. Moreover, the heat losses depend on
current efficiency and energy consumption. Strong velocity fields or cell
oscillations change the current efficiency. Let us analyze such effects by
looking at the calculated and measured results.
8.4.3. Applications
Impact of ledge shape on stability

Magnetohydrodynamic and thermal calculations have been performed for
two stationary cases. In the first case the anode cathode distance was at
its normal set point. In the second case the anode cathode distance (ACD)
was lowered such that a strong ridge was formed under the anodes.
Figure 8-15 shows the ledge profile for each case at the center of the long
side of the cell.

8-21
8. Magnetic Effects
(A) B)
Figure 8-15: Ledge Formation. (A) standard ledge, (B) ridge under anode.
The existence of ridge is directly correlated with a decrease of current effi-
ciency and this is due to the loss of cell stability.
(A)
(B)
Figure 8-16: Horizontal component of the velocity field in the middle of the metal pad.
(A) standard ledge corresponding to Figure 8-15(A) (maximal velocity
0.21 m/s). (B) ridge corresponding to Figure 8-15(B) (maximal velocity
0.33 m/s).

8-22
8. Magnetic Effects
Figure 8-16 (A) and (B) show the corresponding velocity fields in a hori-
zontal plane at the middle of the metal pad. Both magnitude and orienta-
tion of the velocity field are strongly modified by the ledge change.
(A)
(B)
Figure 8-17: Represents the real (A) and imaginary parts (B) of the wave circular fre-
quencies in a complex plane.
Once the stationary solutions are known, one can solve Equation 8-32 to
8-42 numerically and find out about the stability of the cell. Figure 8-17 (A)

8-23
8. Magnetic Effects
and (B) represent the real and imaginary parts of the wave circular fre-
quencies in a complex plane.
The small triangles represent the position of each mode when there is no
current in the cell. It corresponds to the gravitational modes. When the
current is increased to its nominal value of 125 kA each mode is slightly
displaced on the real axis. This means that the oscillation period is slightly
modified. More important is the fact that the circular frequencies are dis-
placed on the imaginary axis. If the displacement happens to be on the
negative side of the axis, it means a growth factor for the corresponding
wave, and the cell becomes unstable. If the displacement is on the positive
side of the axis, it means a dumping factor, the cell remains stable.
With the assumed operational parameters, two modes are slightly unsta-
ble. Mode (3,0) and (3,1) have respectively a growth factor of 1.2 10-3 s-1
and 2.7 10-3 s-1. When the ledge moves under the anodes, most of the
modes are unstable (Fig. 8-17B), and the growth factors are increased.
Mode (1,0), with a very long oscillation period of 75 s, was absolutely sta-
ble when the ledge was good. With a ridge, it is the most unstable mode
with a growth factor of 6.6 10-3 s-1. Three other modes have a growth fac-
tor bigger than 5.0 10-3 s-1. Under such a state action must be taken in
orded to stabilizing the cell. One solution is to increase the anode cathode
distance.
Impact of anode cathode distance

Let us consider the case when the ACD has changed and the cell has not
yet reacted with a new ledge profile. Figure 8-18 shows the impact when
decreasing the ACD from 5.8 cm to 3.8 cm. In this case three modes are
very unstable. Modes (1,0), (2,0) and (0,1) have a growth factor bigger
than 1.3 10-2 s-1 which is much greater than observed with a strong ridge.
It confirms the strong impact of the ACD on the cell stability. It can also be
observed from the figures that the oscillation periods are modified by the
ACD. In case of mode (1,0) the oscillation period changes from 75 sec-
onds to 83 seconds.

8-24
8. Magnetic Effects
Figure 8-18: Real and imaginary parts of wave circular frequencies. Anode cathode dis-
tance reduced to 3.8 cm.
The same type of analysis can be performed when changing operational
parameters. Let us just mention that the metal height is an important factor
for the cell stability. Of course the stability is decreasing when the metal
height is decreasing. The bath level has only a small effect on the stability
due to the pressure field and the velocity field in the channels between the
anodes. The cell is more stable when the bath level is higher.
Busbar system
The busbar system should be designed to give well-balanced magnetic
fields. The optimization of the busbar design, which has a major impact on
current efficiency and cell stability, is based on sophisticated simulation
tools. The use of simulation tools has allowed to optimize both thermal
balance and magnetohydrodynamic states of the cells. In Figure 8-19 the
MHD stability, expressed as a function of the current load, of the original
busbar design and of the modified design are compared. Keep in mind that
the growth factor is the reciprocal time to double the amplitude of the cor-
responding eigen-mode. Hence, big values indicate a tendency to build up
instabilities. Busbar modifications show the stabilizing effect on the mag-
netohydrodynamic state of the cell. It allows to further increase of the cur-
rent load.

8-25
8. Magnetic Effects
Figure 8-19: MHD stability comparison versus current load for original and optimized
busbars.

8-26
8. Magnetic Effects
8.5. Literature
Lit. 8-1: V.F. Almukhametov et al.

“Magneto-Hydrodynamic Phenomena in Production of Aluminium by
Electrolysis”
Light Metals 1990, pp. 249-256.
Lit. 8-2: R. Shekhar, J.W. Evans

“Modeling Studies of Electrolyte Flow and Bubble Behavior in Ad-
vanced Hall Cells”
Lit. 8-3: R. Boivin, S. Martel,

“Effect of an Instability of the Metal Surface on the Magnetic Field In-
side a Cell”
Light Metals 1990, pp.233-241.
Lit. 8-4: D.C. Chesonis, A.F. LaCamera,

“The Influence of Gas-Driven Circulation on Alumina Distribution and
Interface Motion in a Hall-Héroult Cell”
Lit. 8-5: J.M. Purdie, et al.

“Impact of Anode Gas Evolution on Electrolyte Flow and Mixing in Alu-
minium Electrowinning Cells”
Lit. 8-6: D. Vogelsang, M. Segatz, C. Droste, P. Baekler and R. Stücher

“Development of a 300 kA Reduction Cell: Application of Simulation
Tools for the Conceptual Design”
Lit. 8-7: A.I. Begunov, B.S. Gromov

“Fluid Dynamic Effects of Gas Circulation in Aluminium Cells”
Lit. 8-8: M. Segatz, D. Vogelsang, C. Droste and P. Baekler

“Modelling of Transient Magneto-Hydrodynamic Phenomena in Hall-
Héroult Cells”
Lit. 8-9: M. Segatz, C. Droste,

“Analysis of Magnetohydrodynamic Instabilities in Aluminium Reduc-
tion Cells”
Lit. 8-10: M.M. Bilek, W.D. Zhang, F.J. Stevens

“Modelling of Electrolyte Flow and its Related Transport Processes in
Aluminium Reduction Cells”

8-27
8. Magnetic Effects
Lit. 8-11: Mori K., Shiota K., Urata N., Ikeuchi H.

“The Surface Oscillation of Liquid Metal in Aluminium Reduction Cells”
Light Metals 1976, Vol. 1 pp. 77-95.
Lit. 8-12: Urata N., Mori K., Ikeuchi H.

“Behavior of Bath and Molted Metal in Aluminium Electrolytic Cell”
Keikinzoku, 26(11), 1976, pp. 573-583.
Lit. 8-13: Sele T.

“Instabilities of the Metal Surface in Electrolytic Cells”
Light Metals 1977, Vol. 1 pp. 7-24.
Lit. 8-14: J.P. Antille, M. Givord, Y. Krähenbühl, R. Von Kaenel

“Effects of Current Increase on Aluminium Reduction Cells”
Lit. 8-15: M.V. Romerio, J. Descloux, M. Flück

“Linear Stability of Electrolysis Cells Parts I,II”
EPFL, DMA, Novembre (1996).
Lit. 8-16: M.V. Romerio, J. Descloux, M. Flück

“Stability Analysis of an Electrolytic Cell for Aluminium Production by a
Perturbation Method Parts I,II”
EPFL, DMA, Septembre (1991).
Lit. 8-17: J. Descloux, Y. Jaccard, M.V. Romerio

“Bidimensional Stability Results for Aluminium Electrolytic Cells”,
January (1991),
Journal of Computational and Applied Mathematics 38, (1991) 1, pp. 77-85.
Lit. 8-18: V. Bojarevics, M.V. Romerio

“Long Waves Instability of Liquid Metal-Electrolyte Interface in Alumin-
ium Electrolysis Cells: a Generalization of Sele’s Criterion”
European Journal of Mechanics, B/Fluids, Vol. 13, N°1,33-56, (1994).

8-28
9. Raw Materials
9. Raw Materials
9.1. Summary ...................................................................... 9-3
9.2. Alumina ........................................................................ 9-3
9.3. Anodes ......................................................................... 9-5
9.4. Cryolite ......................................................................... 9-9
9.5. Aluminum Fluoride.................................................... 9-10
9.6. Sodium Carbonate .................................................... 9-11
9.7. Literature.................................................................... 9-12

9-1
9. Raw Materials

9-2
9. Raw Materials
9. Raw Materials
9.1. Summary
The composition and the physical properties of the following raw ma-
terials are specified: alumina, carbon anodes, cryolite, aluminum
fluoride and sodium carbonate.
9.2. Alumina
The following are approximate data for the alumina used for problem-free
operation of point feeding pots with dry scrubbing.
Table 9-1: Specification of the alumina composition (alumina dried at 120 °C).
Compound Limiting Concentration
Alumina (Al2O3) > 98 %

Sodium Oxide (Na2O) < 0.4 %
Loss of Ignition (LOI) 0.4 to 0.8 %
Silicium (Si) < 100* ppm
Iron (Fe) < 100* ppm
Zinc (Zn) < 100* ppm
Titanium (Ti) < 20* ppm
Phosphorus (P) < 100 ppm
Vanadium (V) < 20* ppm
Calcium and Magnesium (Ca, Mg) < 500 ppm
Gallium (Ga) < 100* ppm
*these limiting values depend on the purity required by the metal.
Determination of Loss of Ignition (LOI):

The sample is dried at 300 °C, heated up to 1000 °C and the left to cool in
a desiccator. The LOI value is the loss of mass of the sample.
At a high degree of pot gas recovery, contents of sodium oxide higher than
0.4% result in production of cryolite (see Exercise 2-6: Sodium Oxide in
Alumina).

9-3
9. Raw Materials
Table 9-2: Specification of some physical properties of alumina.
Property Limiting Values
BET Surface Area 70 to 80 m2/g

Density 3.4 to 3.6 g/cm3
α Alumina Content < 20 %
Portion Finer than 45µm <6 %
Determination of the BET Surface Area:

The determination of the specific surface area of alumina uses the adsorp-
tion of gaseous nitrogen at the temperature of liquid nitrogen (-195 °C).
Brunauer, Emmett and Teller (Lit. 9-1) developed the theory for multilayer
adsorption and derived the corresponding equation which is now called the
BET equation.
Effects of Property Changes

Practice has shown that the properties of alumina change during transport
in the dense phase conveying system and in the scrubbing unit.
As a rule, the portion of grains finer than 45 µm increases, which detracts
from the flow characteristics. In an extreme case, this can mean that the
alumina will not flow from the pot hopper into the dosing device.
This can be indicated by very high numbers of overfeeding phases and a
relatively large number of anode effects.
As a counter measure, as little fresh alumina as possible should be used
in electrolysis. This shortens the recirculation time in the scrubber and cuts
the increase of the fine portion. The flow characteristics are also improved
as a result of the reduced fluoride load.
It is a mistake to change in the alumina model the intervals of supply-

ing alumina. If that is done, the system goes out of control!
The moisture content of the alumina in the charge should not exceed
1.3%. Observations have shown that the air burn on the surface of the an-
ode is greater when the moisture content of the alumina is high.

9-4
9. Raw Materials
9.3. Anodes
Table 9-3: Specification of average values of impurity concentrations in anodes.

Concentration
Compound
Limit Typical
Silicium (Si) 400* 100 ppm
Iron (Fe) 900* 400 ppm
Sulphur (S) 2.5 1.5 %
Vanadium (V) and Nickel (Ni) 400* 250 ppm
Natrium (Na) 0.12 0.06 %
Ash 1.5 1.0 %
The limit value for sulphur is determined by the acceptable limits for sul-
phur dioxide emission and/or immission and by the sulphur content of
alumina.
Table 9-4: Specification of physical properties of anodes.

Value
Property
Limit Typi-
cal
Apparent density ≥ 1.50 (L-2σ) 1.58 g/cm3
≤ 1.61 (L+2σ)
Spec. elec. Resistance ≤ 70 (L+2σ) 58 μΩm
Bending Rupture Strength ≥ 80 110 105 N/m2
Density in Xylol ≥ 2.05 (L-2σ) 2.07 g/cm3
Reactivity in CO2: Rest < 83 ~ 90 %
Reactivity in CO2: Dust <6 %
Reactivity in Air: Rest < 83 > 90 %
Permeability <2 nPm
Determination of the Specific Electrical Resistance of anodes:

The specific electrical resistance is determined by measuring the voltage
drop on a 50 mm diameter by 130 mm long sample at a constant current
of 1 ampere. The test arrangement is shown in Figure 9-1. The results are
expressed in micro-ohm meters.

9-5
9. Raw Materials
Figure 9-1: Test arrangement for the determination of the specific electric resistance of
anodes.
Determination of the Flexural Strength of anodes:

The flexural strength is determined by using three point loading on a sam-
ple with a diameter of 50 mm and a length of 130 mm. The test arrange-
ment is shown in Figure 9-2.
Figure 9-2: Test arrangement for the determination of the flexural strength of anodes.
Determination of the Carboxy Reactivity of anodes:

The carboxy reactivity is tested by heating a sample of 50 mm diameter
and 60 mm length for 7 hours at 960 °C in a carbon dioxide atmosphere
(Figure 9-3). The selective carbon dioxide burn is measured by placing the
sample after the test together with steel balls for 20 minutes in a tumbling
apparatus and then weighing the removed dust.

9-6
9. Raw Materials
Figure 9-3: Test arrangement for the determination of the CO2 reacticity of anodes.
Determination of the Air Reactivity of anodes:

Figure 9-4 shows the test arrangement to determine the air reactivity be-
haviour of anodes. The sample (diameter 50 mm and 60 mm length) is
preheated in an inert atmosphere at 550 °C and then cooled with a gradi-
ent of 15 °C/h to 400 °C in an air flow of 200 I/h. In order to collect the dust
in a cold area the sample is cyclically tapped.
Figure 9-4: Test arrangement for the determination of the air reacticity of anodes.
Determination of the Air Permeability of anodes:

The air permeability is determined by measuring the time taken for a cer-
tain volume of air to pass through a sample of defined dimensions. The
results are expressed in nanoperms (1 nPm = 1 Darcy/9.87).

9-7
9. Raw Materials
Table 9-5: Specification of anode dimensions.
Dimensions Limiting Values
Length ± 10 mm
Width ± 6.25 mm
Height ±8 mm
Deformation Long Side < 40 mm
Deformation Bottom Side < 30 mm
If, for any reason, anodes have to be used that do not conform to these
specifications, they should only be used "thinned out." For each pot, there
should never be more than one anode in the anode set that does not con-
form to specifications.

9-8
9. Raw Materials
9.4. Cryolite
Table 9-6: Specification of the cryolite composition.
Fluorine ≥ 53 %
Calcium Fluoride (CaF2) ≤5 %
Cryolite Ratio ≥ 2.2
Sulfate Ion (SO-24) ≤ 0.25 %
Silicium Oxide (SiO2) ≤ 0.1* %
Iron Oxide (Fe2O3) ≤ 0.1* %
Phosphorus (P) ≤ 100 ppm
Moisture ≤ 0.3 %
To get information about the cryolite ratio see its definition (Equation 2-10
in Chapter 2-3).
The grain size of the cryolite should be less than 2 millimetres.

9-9
9. Raw Materials
9.5. Aluminum Fluoride
Table 9-7: Specification of the aluminum fluoride composition.
Fluorine ≥ 60 %
Aluminum Fluoride (AlF3) ≥ 90 %
Calcium Fluoride (CaF2) ≤2 %
Magnesium Fluoride (MgF2) ≤ 0.1 %
Phosphorus (P) ≤ 100 ppm
Loss of Ignition (LOI) ≤ 0.9 %
The grain size of the aluminum fluoride should be less than 2 millimetres.

9-10
9. Raw Materials
9.6. Sodium Carbonate

Use only calcined sodium carbonate !
Table 9-8: Specification of the sodium carbonate composition.
Sodium Carbonate (Na2CO3) ≥ 98 %

Chlorine Ion (Cl-) ≤ 0.3 %
Calcium Oxide and Magne- ≤ 0.2 %
sium Oxide (CaO, MgO)
Humidity ≤ 0.2 %
When heated up sodium carbonate generates a great amount of carbon

dioxide. For safety reasons the use of pelletized soda is recommended
due to this extremely strong generation of carbon dioxide. The formation of
gas takes then place over a more extended period of time and is thus sig-
nificantly less brisk.
Possible source of sodium carbonate: Deutsche Solvay AG.

9-11
9. Raw Materials
9.7. Literature
Lit. 9-1: S. Brunauer, P. H. Emmett, E. Teller,
"Adsorption of Gases in Multimolecular Layers",
J. Am. Chem. Soc. (1938) 60, 309-319.

9-12
10. Cathode Lining and

Pot Startup
10.1. Summary................................................................. 10-3
10.2. Cathode Lining ....................................................... 10-3

10.2.1. Introduction............................................................................................... 10-3
10.2.2. General..................................................................................................... 10-4
10.2.3. Insulation of the Shell Floor...................................................................... 10-4
10.2.4. Casting-in the Bars................................................................................... 10-6
10.2.5. Insertion of the Bottom Blocks ................................................................. 10-8
10.2.6. Sealing Off the Bar Windows ................................................................. 10-11
10.2.7. Lining the Shell Walls ............................................................................. 10-12
10.3. Start Up ................................................................. 10-16

10.3.1. General................................................................................................... 10-16
10.3.2. Electrical Calcination .............................................................................. 10-17
10.3.3. Gas Calcination ...................................................................................... 10-20
10.3.4. Bathing Up.............................................................................................. 10-22
10.3.5. Normalization Phase .............................................................................. 10-24

10-1

10-2
10. Cathode Lining and Pot Start-Up
10.1. Summary
This chapter describes the steps to make the Cathode Lining

namely: insulating the shell floor, casting-in of the bars, insertion
of the bottom blocks, sealing off the bar windows and lining the
shell side walls.
The procedures to Start Up an electrolytic pot are given: electric
calcination, gas calcination, bathing up, nomalization phase.
10.2. Cathode Lining
10.2.1. Introduction
Did you know
that properly directed added expense and effort in repairing cathodes

can prolong the potlife and thereby result in large savings?
that cathodes of the most modern design have life expectancies of up
to 3000 days?
that these expectancies can only be realized if the cathodes are lined
and put into operation using painstaking care and materials of the
highest quality?
that extensive tests are performed, looking for and finding new meth-
ods and means to improve the quality of the cathodes and prolong the
potlife?
that it is up to you to utilize the results of these tests to your advan-
tage?
that other people are interested in the experience you are gaining?
Naturally you know all that – but are you also turning that knowl-
edge into real value?
The more exact the data collected about a cathode, the more you can
learn from your experience with it. Please help us in our gathering of data
so as to create a basis for further development.
The text that follows has been based on cathodes with Type C shells.
With the aid of the inclosed drawings, all work instructions and data
provided here can also be transferred to other cathodes types.

10-3
10.2.2. General
A high quality cathode ensuring long potlife is characterized by a uniform
current distribution and a lining that metal and the electrolyte cannot erode
and penetrate.
Type C cathodes have been designed in accordance with most recent
knowledge so as to satisfy these requirements. The cathode takes up
stress-free the differing thermal expansions of the materials that are un-
avoidable while being heated to operating temperature. This avoids uncon-
trolled changes in shape and cracking during the start-up, both of which
severely reduce the quality of the cathode.
The great advantages of this design can attain their full effect on the be-
havior of the pot only if pots are carefully lined.
Figure 10-1: Steel shell with cathode lining.
10.2.3. Insulation of the Shell Floor

The shell lining serves as thermal insulation and as basement for the
cathode blocks. Its components must be resistant to electrolyte and alumi-
num at a temperature of 950 °C or must be protected against their pene-
tration.
The main components of the lining are:
Firebricks: Resistant to electrolyte and liquid aluminum.

Vapor barrier: The vapor barrier is made of sheet steel to prevents so-
dium compounds and aluminum from penetrating into
the layers of insulation underneath.
Insulating the insulating bricks together with the plates and gran-
bricks: ules are not or only slightly resistant to electrolyte and
molten aluminum.
Figure 10-2: Schematic structure of cell lining.

10-4
Insulating Layers
Two inslation layers are laid in loose. The first layer consists of insulating
plates. If graphite blocks are used the second layer consists of insulating
plates in the edge region and bricks in the center. When using amorphous
blocks, only insulating bricks are layed in the second layer.
Figure 10-3: Laying the insulating plates and bricks.
Vapor Barrier
The vapor barrier consists of 4 layers of sheet steel (St 37), 1 mm thick.
The individual sheets cover the entire width of the shell and shoved up
loose against one another (not welded!). The sheets in the upper layer
overlap those underneath by 50% in each case.
As an alternative to this, the vapor barrier can also be built from two steel
plates of appropriate dimensions, 4 mm thick, welded together gas-tight
after being laid in place. Any excess welded seam has to be ground flat!
Firebricks
Two layers of firebricks are laid in loose. The joints between the bricks are
filled with fine grained fire-clay or alumina. In the third layer fire-clay mortar
is used.
The top surface of the firebrick must be 10 mm higher than the bot-
tom edge of the bar window!

10-5
Figure 10-4: Height difference between cell lining and bar window..
After lining the mortar must dry for at least 16 hours using an infrared
heating cover (up to 110 – 130 °C).
10.2.4. Casting-in the Bars
The cathode bars are fastened in the grooves of the cathode blocks by
casting them in with iron.
Figure 10-5: Finished cathode element.

The cast fastening must satisfy the following requirements:
• The transition resistance between the carbon block and the cathode
bars should be minimal.
• No tensile stresses greater than the tensile strength may be produced
in the carbon either during casting-in or while heating to operating
temperature. Such stresses would cause cracking.
The following work instructions have been based on calculations made for
casting-in of double bars in amorphous cathode blocks. As graphitic cath-
ode blocks behave thermomechanically better than amorphous blocks,
and as stresses produced when casting-in single bars are less than those
produced when casting-in double bars, these work instructions are valid
for graphitic cathode blocks and single bars as well.

10-6
The high quality of cathodes and long potlife are only guaranted if
the work instructions are strictly followed.
Preheating before Casting-in

The preheating temperatures have been defined such that no tensile
stresses exceeding the tensile strength of the carbon will occur during
casting-in. This avoids cracking during the casting-in process. Mechanical
pre-stressing of the cathode blocks can also help to prevent excessive
stresses.
Figure 10-6: Temperature measurement points on cathode element.

Prescribed temperature:
Cathode bars: 650 - 700 °C,
Carbon, outer surface: 350 - 400 °C,
Carbon, close to bars: 430 - 480 °C.
Figure 10-7: Preheating device for cathode element.
Production of the Cast Iron

Temperature and composition control of the cast iron are of particular im-
portance.
Uniformity of the properties and/or the cast structure of the cast iron are
factors determining the quality of the cathode and its life expectancy be-
cause they determine how the bottom blocks will behave when being pre-
heated to operating temperature. It is therefore extremely important that
variations in cast iron quality are as small as possible. In addition to the

10-7
chemical composition, the cast iron solidification process must be con-

trolled. Depending on its graphitization, cast iron melt solidifies either
white, whereby it shrinks approximately 4%, or gray, whereby it expands.
In practice, the melts solidify somewhere between the two extremes, and
the volume changes correspondingly.
When being heated to operating temperature, the mechanical stresses oc-
curring due to differences in the graphitization process may be changed
correspondingly. The tensile stresses in the carbon may exceed permissi-
ble limits and cracks are inevitable.
The cast iron is molten in an induction or a rotary furnace.
The overheating (i.e., heating to temperatures above the melting point)
and its duration should be kept to the possible minimum, i.e., the cast
iron is made ready at a predetermined time. Calculate the time for
charging the furnace based on the melting capacity of the furnace and the
needed iron amount.
Table 10-1: Composition range of elements in cast iron.
The saturation degree (SC) of carbon in cast iron should be 1.0 to 1.1.
cC
SC =
d
4.3 − 0.33 × c Si + c P i (10-1)
10.2.5. Insertion of the Bottom Blocks

Glueing the bottom blocks has the following Advantages:
• less working time is required for lining the pot,

• the working place atmosphere is improved during pot lining,
• it excludes quality variations from ramming.
The cathode blocks used for glueing must be machined exactly. The cra-
dle shell and the lining have been designed such that they can take up the
thermal expansion of the entire glued cathode stress-free.

10-8
Work Routine
• Determine the position of the first bottom block at the center in such a
way that the bars are centered exactly at the corresponding bar win-
dows.
• Cover the inner collar of the bar on the side facing the wall of the shell
with Fiberfrax coated with graphite powder. Pass it over the ends of
the bar.
• With a crane, lift the middle bottom block into the shell.
• Introduce the support beam and brace it against the closer wall.
Figure 10-8: First bottom block in position.

• Put the second bottom block on the edge of the shell above the mid-
dle bottom block with a crane.
• Before doing so, pass the inner bar collars over the ends of the bars.
• Coat the long sides of the two bottom blocks that are to be glued to-
gether with a thin coating (max. 1 mm) of glue leaving an edge at the
bottom of approximately 30 mm uncoated.
• Lift the second bottom block with the crane. Turn it 180° and set in
down into the shell next to the first element.
• Measure the difference in height between the second and the first
element.
Figure 10-9: Positioning of the second bottom block.

The acceptable difference in height for all glued cathode blocks with re-
spect to the first inserted element is (+/-) 3 mm.

10-9
Figure 10-10: Tolerances for inserting bottom blocks.
• Press the two cathode blocks against one another with hydraulic
presses.
The hydraulic pressure exerted on the glued surfaces should be
0.05 to 0.1 N/mm2.
• Press the ends of the bars on the two cathode blocks together with
clamps on both sides. Tighten the clamps with a torque wrench.
A tightening torque is selected such that the pressure exerted on the

surfaces of the carbon will be 0.05 to 0.1 N/mm2 when the hydraulic
presses are removed later.
Figure 10-11: Bracing the bottom blocks.
• Remove any excess glue from the surfaces of the cathode blocks.
• Insert the third and all further cathode blocks into the shell and glue
them together as described above.
• Measure and record the ambient temperature.
Select a hardening time such that a minimum bending strength of 10

N/mm2 is attained. Minimum hardening time (valid for SGL-glue) :
98 h at 10 °C
27 h at 20 °C
9 h at 30 °C
9 h at 40 °C.
• After hardening remove the clamps from the ends of the bars.

10-10
10.2.6. Sealing Off the Bar Windows
Sealing Off the Bar Window from Inside
Figure 10-12:Sealing the bar window.

In the inner bar collar four threaded holes are drilled for the screws. As
tightening material Fiberfrax with graphite is used.
Figure 10-13: Wrap up the bar with Fiberfrax rope.

Inner and outer bar collar are pulled together by screws so that their
flanges press against the shell wall (torque: 12 Nm). Fiberfrax coated with
graphite powder between flanges and shell allows the collar to glide on the
walls when the carbon bottom elongates during preheating.
Sealing Off the Bar Window from Outside

• The gaps between collars and iron bars are plugged up with ropes.
• Then the torque of the collar screws is increased to 43 Nm.
• By this design the iron bar can move in the windows in longitudinal
and transverse direction as well.
• After four weeks in operation the screws must be re-tightened. In ad-
dition should be checked that the seal is tight. If necessary, the seal
should be repaired.

10-11
Figure 10-14: Sealing the bar window from outside.
10.2.7. Lining the Shell Walls
Filling in Concrete
The spaces between bars, corners and traverse sides are filled with con-
crete. Starting from the outside and going inward, the lining of the shell
walls consists of side and corner blocks and ramming paste. Due to the
expansion gaps between the concrete and the bottom blocks, only minimal
thermal stresses occur during calcination.
• Before pouring the concrete, place Styropor plates 15 mm thick

against each bar on the side facing the closest transverse side.
• The Styropor plates and the expansion gaps allow stress-free thermal
expansion of the cathode.
Figure 10-15: Concreting corners, long and transvers side of the pot.

10-12
Figure 10-16: Structure of the shell wall.
Insertion of the Sidewall Blocks
Figure 10-17: Insertion of sidewall blocks.

The last sidewall block must be fitted in.
Ramming Procedure
The ramming paste becomes completely hard only at temperatures of
more than approximately 350 °C. This means that it can still compensate
partially the differences in the thermal expansion of the various materials
in the shell lining during calcination, thereby preventing mechanical
stresses. As the ramming paste also hardens at room temperature –
though very slowly – the pot should be put into operation as soon as pos-
sible after the ramming procedure.
Ramming is done either by a group of workers using compressed-air
hammers or vibrators, or machines with varying degrees of automation.
The compaction of the ramming paste is only uniform and reproducible

10-13
when machines are used. However, the corners always have to be

rammed manually. If a ramming paste is compacted too much or too long
(particularly using compressed-air hammers), or if the ramming tempera-
ture is higher than the range indicated by the manufacturer, as a result oc-
cur laminations during pot operation.
When using cold ramming paste, the temperature of the cathode

must be at least 18 to 20 °C during the ramming procedure. This
means that in unheated rooms, preheating is necessary during winter. Af-
ter the lining has been dried, it is necessary to wait until the cathode tem-
perature has dropped below 30 °C before starting the ramming procedure.
After ramming, the ramming paste must have a strength of 3.5 N/mm2.
This strength is determined with a penetrometer which measures the resis-
tance to penetration.
Ramming the Expansion Gap
Figure 10-18: Rammed layers in the finished expansion gap on the short side of the pot.
Lining the Side Walls with Ramming Paste

Seven layers are rammed so that their surfaces drop off towards the cath-
ode block.
The seventh layer should be designed in a way that it starts 50 mm below
the edge of the cathode blocks and covers over it by a distance of 50 mm.

10-14
Figure 10-19: Ramming layers on the finished side wall.
Figure 10-20: Fully rammed corner.

10-15
10.3. Start Up
10.3.1. General
There are three phases in starting up an electrolysis pot: calcination, bath-
ing-up, and the normalization phase. The total process lasts 4 to 5 weeks.
During calcination and bathing-up, the cathode is heated to operating tem-
perature. If the cathode is put into position and pre-heated for bathing-up
as prescribed, the thermal stresses remain below the critical limit even
when temperature gradients occur, therefore the various materials expand
differently. Cracking can therefore be prevented.
The collector bars have the biggest thermal expansion. They can move
relatively free in the groove up to a temperature of approx. 600 °C. Above
these temperatures, a friction bond forms and tensile stresses are pro-
duced in the cathode block. Heating the cathode from the center and
working outward ensures that the friction bond will start at the center and
that the bars can expand freely outward for as long as possible. An early
friction bond at two locations on the collector bar produces tensile stresses
so high that the carbon can crack.
As the carbon ramming paste begins to harden during calcination at tem-
peratures of more than 300 °C, it is important that the base of the cathode
is heated at the center first so that it can expand as long as the ramming
paste is still plastic.
The goal is to obtain a preheating in which the carbon ramming paste
hardens from the bottom. This can, to a large extent, prevent lamination of
the ramming paste within the rim.
Preheating is the first step of the start-up. There are two possible proce-
dures available for accomplishing this:
• electrical calcination for new cathodes,

• gas calcination for cathodes that have been shut down temporarily.
During electric calcination, the line current flows through the pot and sup-
plies energy for heating and calcination. To keep the temperature rise flat
enough during the initial phase, a shunt reduces the pot current initially. An
uniform anodic current distribution is important for good electrical calcin-
ing.
For pots up to ~110 kA, with single anodes and with glued cathode blocks,
electrical calcination without a shunt is also feasible, but the risk of local
overheating is significantly greater than when using a shunt.
For gas calcination, the pot is heated and calcined with gas burners placed
between the anodes and the cathode.

10-16
10.3.2. Electrical Calcination

For balanced calcining temperature across the entire pot, the anodic cur-
rent distribution must be as uniform as possible during calcining. Varia-
tions in the anode currents due to differences in the level of the coke bed
and/or unequal contact between the anodes and the coke lying under
them can be avoided to a large extent by carefully preparing the pot.
Preparations:
• In the anode rodding, select crack-free anodes with bottoms that do
not deviate by more than 4 mm from exact surface evenness
• Brush the anode rod clean especially at the contact points to the
flexible connections.
• Position the anode beam on the pot at 25 cm and check all equip-
ment.
• Install the flexible anode connections on the anode beam.
• Distribute the coke granules (1.0 - 3.0 mm) on the cathode surface
with a leveling bar, level the layer to a thickness of 10 - 12 mm.
• Clean the bottom of the anodes.
• Introduce the first anode into the clamp at a distance of 1 - 2 cm from
the anode beam.
• Install the flexible connection on the anode stem. Do not tighten the
clamp.
• Raise the anode again slightly and press it on the coke layer.
• Check that the anode rod is at a sufficient distance from the horizontal
covering.
• Tighten the flexible strips on the rod and on the anode beam.
The anode is now connected to the beam flexibly and can conduct current. It
will travel along with the thermal expansion of the cathode and thereby
maintain the uniform anodic current distribution during calcination.
• Place all other anodes as described above.

• Remove the coke from the spaces between the anodes with a vac-
uum cleaner, so that less carbon dust has to be removed from the
electrolyte during the first days of operation.
• Attach angled steel plates (0.7 mm thick) to the outsides of the an-
odes at long and short sides of the pot.
• Mix 100 kg of soda with 5 to 6 parts of cryolite or crushed bath and fill
the mixture into the space between the anodes and the side blocks.
The soda should be evenly distributed to prevent explosions. In con-

tact with liquid bath 1 kg of soda produces approximately 200 l of
CO2 gas. The use of "briquetted" soda (50 x 30 x 25 mm, from Deut-
sche Solvay) has worked out well. Gas development takes place
more slowly (see Chap. 9-6 and Exercise 2-7).

10-17
The needed soda amount depends on the pot size and the amount of ex-
cess aluminum fluoride in the electrolyte. The electrolyte in the starting pot
must be brought to an excess aluminum fluoride level of 0-3% within the
first 24 hours (electrolyte sample is light pink).
Calcination:
Figure 10-21: Installation of shunts.
• Fill the remaining space with crushed bath.

• Put sheet metal strips across the gaps between the anodes. Use
plates with holes for temperature measurement.
• Put the thermocouples into position and connect them to the display
instrument.
• Lay insulating pads on top of the metal strips.
• Before mounting the shunts in the basement clean the contact points.
Connect the shunt to all bus bars (see Figure 10-21).
• Bring the pneumatic short-circuiting devices into position and start
them up.
• Take off the short-circuiting plates.
• Remove the pneumatic short-circuiting device.
• Remove the insulating short-circuit of the preceding pot.
• Check contact resistances of the shunt. Improve any contacts that
have a voltage drop bigger than 10 mV by wedging them more
strongly.
• Measure the current of each anode. If anodes draw no current, rap
them, to press them better against the coke bed.
At this point it is not possible to make a reasonable measurement of

the actual current distribution since the distribution is still very much in-
fluenced by the contact between the anode rod and the anode block.
• After heating up for two hours, measure and record the anodic current
distribution.

10-18
During the entire calcination process, the Variation Coefficient (VK Equa-
tion 10-2) of the anodic current distribution should be less than 15%.
σ (10-2)
VK = ⋅ 100.0
x
where the Standard Deviation is determined according to Equation 10-3
n
∑ ex j
2
i −x (10-3)
σ= i =1
n
where the Mean Value according to Equation 10-4.

n (10-4)
∑ xi
x = i =1
n
The current distribution can be improved by striking or rapping on the an-

ode rods that are drawing too little current or by tightening more firmly the
screw clamps fastening the contact plates of the anode flexible. It is also
recommended to remove the insulation pads of anodes which are drawing
too much current.
• Measure the anodic current distribution every 4 hours and record the
variation coefficient.
• After 4 hours, or when the pot voltage has dropped below 3.0 V, re-
move one pair of shunts from the electrical circuit.
• After 8, 12, 16, and 20 hours, or every time the pot voltage drops be-
low 3.0 V again, remove one more pair of shunts.
• after 12 hours, remove the insulation mat over the center of the pot.
• Remove the shunt after 20 hours at the latest.
• Continue calcining at full current for another 28 hours.

10-19
The electrical resistance of the carbon materials drops as the temperature

rises. The shunts keep the energy input to the pot approximately constant,
which results in a uniform temperature curve.
Figure 10-22: Temperature increase during heating up.

After heating for 48 hours, the surface of the cathode blocks should have
reached a temperature of 750 – 850 °C.
10.3.3. Gas Calcination
The burner equipment used for gas calcination should be designed in a
way that the cathode surface can be heated to a temperature of 650 -750
°C within 72 hours.
Preparations:
• Clean the cathode surface carefully.
• Cover the rim and glued joints with sheet metal strips approx. 0.6 mm
thick.
• Bring the anode beam to a height of 15 cm.
• Place the anodes on the floor of the cathode and tighten the clamps.
• Raise the anode beam 25 cm or to its highest position.
• Install the burners between the cathode and the anodes in such a
way that the flames are directed horizontally towards the center of the
pot. Put on insulation covers.
• Cover the gaps between the anodes with metal plates approx. 1 mm
thick. Use plates with holes at the temperature measurement points.
• Bring 6 thermocouples into position and connect them to the display
instrument.
• Light the burners.
Calcination:
During calcining, make sure that a reducing atmosphere prevails between
the anode and the cathode. Ash components on the cathode surface re-
duce the surface tension between the electrolyte and the cathode to zero,
which results in electrolyte penetration and layer forming in the cathode
blocks and/or the ramming paste.
• Calcine the pot for 60 to 72 hours, measuring and recording the tem-
perature of the cathode every two hours. Check at the same time that

10-20
all the burners are in operation.
To ensure that the temperature of the cathode and not that of the gas
flame is being measured, extinguish the burners closest to the ther-
mocouple when taking the measurements.
• Once the calcination temperature has been reached, remove the

insulation plates between the anodes. If the minimum temperature
of 650 °C has not been reached at all measurement points within
72 hours, prolong the calcining time!
• Switch the burners off and remove them quickly.
Carry out the following preparatory work as quickly as possible and without
interruption in order to keep the cooling of the cathode to a minimum:
• Lower the anodes to 1 cm above the cathode. Mark the "0" position.
• At the middle of each long side, keep an opening by using full cryolite
bags on both sides of the hole for pouring in the liquid electrolyte.
Leave the protective plate on the ramming paste.
During gas calcination, the ramming paste is only slightly calcined and
does not reach the same high temperatures as the cathode surface. The
protective plates are not removed because they provide mechanical pro-
tection for the rammed rims which are still soft.
Soda Addition:
• Mix 100 kg of soda with 5 - 6 parts cryolite or crushed bath and fill the
mixture into the spaces between the anodes and the side blocks.
Leave the opening clear for pouring in the liquid electrolyte.
The soda must be distributed to prevent explosions. In contact with

liquid electrolyte, 1 kg of soda produces approx. 200 l of CO2 gas.
"Briquetted" soda (50 x 30 x 25 mm) from Deutsche Solvay has been
used without problems. Gas development takes place more slowly
(see Chap. 9-6 and Exercise 2-7).
The amount of needed soda depends on the size of the pot and the
amount of excess aluminum fluoride in the electrolyte. The electrolyte in
the starting pot must be brought to an excess aluminum fluoride level of 0-
3% within the first 24 hours (electrolyte sample slightly reddish).
• Fill the side gap up to the upper edge of the anodes with crushed
bath.

10-21
10.3.4. Bathing Up
The bathing up are the first 24 hours of pot operation. Electrolytic bathing
of the pot after a gas calcination does not differ significantly from that after
an electric calcination. The electrical calcinations are described below. In
both cases, it is important to introduce the electrolyte as quickly and uni-
formly as possible to prevent overloading and damaging individual anodes.
Table 10-2: Time schedule of the bathing up preparations.
(Time = 0 corresponds of adding the first liquid bath to the pot).
Actual bathing up:
• add 2.5…3.0 t of electrolyte,

• raise the beam to 4 cm,
• the cell voltage will increase to 10…15 V,
• continue to add electrolyte,
• raise the beam to 10…12 cm,
• the electrolyte level should be approximately 25 cm,
• voltage should be between 10 and 15 V,
• cover the anodes with crushed bath and cryolite and a small quantity
of alumina,
• when the bath level of 25 cm is reached, quench the anode effect.
The pot voltage will stay at 6…7 V (approximately 45 min after bath-
ing up),
• the temperature will reach 1000…1030 °C.

10-22
Figure 10-23: Progress of bathing up.

10-23
Table 10-3: Time schedule after bathing up.
(Time = 0 corresponds to adding the first liquid bath to the pot).
10.3.5. Normalization Phase

A four-week normalization phase is used to direct the pot into a thermo-
chemical equilibrium. The metal and electrolyte levels, interpolar distance,
pot voltage, electrolyte composition and pot temperature are brought to
the desired values.
Because the excess of AlF3 in the electrolyte is less than 3% at the start of
the normalization phase, the temperature of the bath must be correspond-
ingly high. The pot is operated initially with a low AlF3 excess and without
metal so that cracks that might have formed during calcination will be filled
with the high-melting electrolyte. This means that such cracks will be sealed
when operating with working temperature of about 960 °C.
During the first 24 hours of operation, the working temperature must be

50-60 °C higher than the given liquidus temperature of the electrolyte. Dur-
ing the normalization phase, the difference between the working tempera-
ture and the liquidus temperature is reduced to 5-7 °C.

10-24
Figure 10-24: Progress of normalization.

Silicon and Iron means the fraction of the silicon and iron concentration
determined on a smaple of the third day. PSP is the Permanent Voltage
Setpoint.

10-25

10-26
11. Process Control
11. Process Control
11.1 Summary................................................................. 11-3
11.2 Philosophy.............................................................. 11-3

11.2.1 Stationary Conditions .......................................................... 11-3
11.2.2 Non-stationary Conditions................................................... 11-4
11.2.3 Process Control Philosophy ................................................ 11-4
11.3 Operating the EPC ................................................. 11-7

11.3.1 Alumina Feed Control ......................................................... 11-7
11.3.2 Alumina Feeding Phases .................................................... 11-7
11.4 Pot Controller Terminal PCT ............................... 11-12

11.4.1 General ............................................................................. 11-12
11.4.2 Manual Entry of Process Events ....................................... 11-13
11.4.3 Display of Pot Data ........................................................... 11-15
11.4.4 Call for Pot Data................................................................ 11-16
11.4.5 Messages.......................................................................... 11-16
11.4.6 Enable/Disable/Testing of Functions/Devices, Data Input. 11-17
11.5 Exercises .............................................................. 11-18

11-1
11. Process Control

11-2
11. Process Control
11. Process Control
11.1 Summary
The programs of the Electrolysis Pot Control (EPC) for pot resis-
tance, noisy pots and under and over feeding initiate the alumina
feed control depending on pot events. They adapt the alumina feed
rate according to the pot state (noisy, quiet). This is done to control
the alumina content in the bath, to avoid unwanted anode effects and
sludge formation. Further this chapter shows how to determine the
permanent voltage set point (PSP) in dependence of bath voltage and
cathode voltage drop.
The meaning of the values shown at the display of the Pot Controller
Terminal (PCT) are explained. The functions of the keys and buttons
are listed. Further the steps to handle events like anode change or
tapping and the way to get data from the PCT are described. The
meaning of the values, messages and alarms shown at the display
are explained in detail.
11.2 Philosophy
11.2.1 Stationary Conditions

Industrial electrolysis pots have many steady states but only one corre-
sponds to the optimum within prevailing economic boundary conditions.
Generally this state is not necessarily characterized by the best technical
performance data as far as for instance the energy consumption is con-
cerned but is defined by operating parameters giving the best economical
results.
As long as those given economic boundary conditions persist, the opti-
mum steady state is characterized by a definite set of operating parame-
ters:
Utot: pot voltage (V),

I: line current (kA),
Ucath: cathode voltage drop (mV),
Ue: external voltage drop (mV),
η × 100: current efficiency (%),
Cb: gross anode consumption (g/kg),
Cn: net anode consumption (g/kg).
In addition it has been shown that bath chemistry can also improve current

11-3
11. Process Control
efficiency. For modern industrial cells the bath composition should have a
high aluminum fluoride excess and a low alumina concentration.
As a result, every set of operating parameters defines an optimum anode-
cathode distance (ACD):
ACD: anode to cathode distance (cm).
It is the task of the operating personnel to maintain the average cell

conditions as close as possible to this optimum.
The various operational routine procedures such as changing anodes,

metal tapping, bridge rolling, anode effect quenching etc., disturb the cell
operation and eventually result in deviations from this optimum.
Industrial cells cannot immediately adapt to operational changes especially
to variations of the heat balance which characterizes each operating pro-
cedure.
11.2.2 Non-stationary Conditions

As already mentioned cells are dynamic due to the inability to adapt im-
mediately to various changes. Therefore cells will deviate from the above
described steady state.
A deviation of one pot parameter from the above mentioned optimum
steady state results in a difference between the actual value and the opti-
mum or target value.
It is the task of Electrolysis Process Control (EPC) to identify this de-

viation and to initiate appropriate action to bring the cell back to the
optimum.
Such a process control can only operate if the overall pot operation phi-
losophy explicitly predetermines the control strategy.
11.2.3 Process Control Philosophy

For cells in operation, where cell design, amperage, anode size and ther-
mal insulation are given, the energy balance is mainly determined by the
pot voltage. To maintain a cell in the ideal steady state the consumption of
the various raw materials has to be compensated, in order to maintain the
energy performance constant.
A cell, however, changes continually its performance and consumption
even if the ACD remains constant at the optimum set point.
With decreasing alumina concentration the resulting decomposition volt-
age (EMF) increases and the ohmic voltage drop decreases:
Cell voltage and cell resistance as function of the alumina concentra-

tion show a minimum at constant ACD.
Figure 11-1 shows the bath voltage as function of alumina concentra tion.

11-4
11. Process Control
Figure 11-1: Bath voltage as function of alumina concentration.
A change of the pot voltage means an undesired change of the energy in-
put of the cell.
Many EPC systems control the pots on a constant voltage basis by adjust-
ing permanently the ACD. This is not the best way to achieve a constant
energy input and this is not the strategy of the ALUSUISSE electrolysis
process control.
Even at constant ACD the energy consumption of a cell changes:
By decreasing the alumina concentration the current efficiency de-

creases, all other parameters remaining constant. This means that
the quantity of effectively produced metal decreases.
If the energy balance remains unchanged decreasing alumina consump-

tion and metal production means that the existing excess energy heats up
the bath:
By increasing the bath temperature the current efficiency decreases

even further and the protective side ledge dissolves.
By decreasing the ACD accordingly to compensate for less consumption it

is possible to reduce or eliminate the excess power but without correcting
the depletion of the alumina content in the bath.
The basic functions of the ALUSUISSE electrolysis process control phi-
losophy are the following:

11-5
11. Process Control
• The anode cathode distance is kept constant.

• A decrease of the alumina concentration is precisely determined
by the EPC and adjusted by adding alumina so that the bath volt-
age remains nearly constant.
By this means current efficiency, bath temperature and energy input re-
main constant.
Therefore it is not necessary to adjust the cell power for short term
variations of the cell consumption.
In order to keep the alumina concentration as constant as possible, small

variations must be admitted in order to determine over- and underfeeding
cycles. However, they must be so small, that despite the constant ACD
hardly excess energy is generated.
The width of the resistance control band allows these small variations of
the alumina concentration without changes of the ACD. A narrow control
band is introduced from time to time, in order to check, whether the ACD is
on the optimum steady setpoint.
The alumina addition through point feeders must be controlled so far, that
over a long period of time the addition is nearly the theoretical consump-
tion. This procedure is only interrupted, if strong underfeeding is necessary
as a control means.
In conclusion the process control philosophy can be summarized as fol-

lows:
The energy input and consumption of the cell are kept constant at a
constant optimum ACD by adjusting the alumina feeding rate to keep
the alumina concentration in the bath within a very narrow concen-
tration band as long as possible.

11-6
11. Process Control
11.3 Operating the EPC
11.3.1 Alumina Feed Control

The control of the alumina feeding for EPT10 pots can be initiated by di-
rect measurement of the pot voltage and the cell current or by manual in-
tervention during usual pot operation procedures (push buttons at the pot
control unit or PCU for tapping, anode change, beam raising and manual
alumina feeding).
Principally three control programs are used for the alumina feed control:
• resistance control: RC
• noisy pot control: NPC
• prediction control: PC (under- and overfeeding)
According to any event or intervention on the pot the control programs are
active or inactive. That means for example:
• if a pot has been identified as noisy and is operated at the highest cor-
rective level, then the over- and underfeeding are inactive because for
this pot in this state the R-slope cannot be calculated accurately. This
situation leads to an unprogrammed anode effect.
• if the R-slope exceeds certain target values the resistance control pro-
gram cannot move the anode beam downwards, even if in this case a
too high resistance has been found.
• if the beam raising button has been activated the anode beam is
blocked for automatic movements even when the resistance is outside
the band width.
11.3.2 Alumina Feeding Phases
Fundamentals
Control of the feeding of alumina is based on the ALUSUISSE pot operat-
ing system conceived for micro processor controlled point feeders:
At constant and optimum ACD a constant energy performance is

achieved by adjusting the alumina feed rate by using the resistance
slope to maintain a nearly constant alumina concentration in the
bath.
The control programs do not react only to every concentration change in

the bath but identify each possible state of the pot, warm and cold, noisy
and quiet, in order to avoid sludge formation as well as unwanted anode
effects.

11-7
11. Process Control
To ensure a proper reaction to any change in the pot and to adjust the
feeding of alumina correspondingly the current identified/calculated pa-
rameters have to be compared to the base parameters.
The Base Parameters can be the followings:
Bath level: 20 cm,

Aluminum fluoride concen- 12 %,
tration:
Average alumina concentra- 2 %,
tion :
Average bath temperature: 960 °C,
Bath mass: 4 - 5 t,
Current efficiency: 93 %,
Line current: 118 kA,
Alumina consumption: 70.7 kg/h.
To keep the alumina concentration constant, 70.7 kg of alumina

should be added to the bath per hour.
You determine the time period [t100] for 100% alumina feed by taking into
account that two point feeders add alumina to the bath with:
2 × 18
. × 60 (11-1)
I 100 = = 3.06 min = 183 sec
70.7
You may also use ElysePrg to determine these values. Click in the List
Examples and Exercises Window of ElysePrg (Chap. 18.14) on Exp. 11-1
and ElysePrg executes the calculation of Exp. 11-1. Figure 11-2 shows the
result of the calculation.

11-8
11. Process Control
Figure 11-2: Result of a calculation the alumina feeding rate using ElysePrg.
Table 11-1 lists the feeding rate (in percent of the theoretical alumina feed-
ing rate) and the duration of the different alumina feeding phases.
Table 11-1: List of the alumina feeding phases.

11-9
11. Process Control
In the following you find a description of the alumina feeding phases:
AE Killing of Anode Effect

When the EPC detects an anode effect, this phase is initiated. The time
between the occurrence of the anode effect and the start of the alumina
feed can be up to 15 sec.
This phase cannot be interrupted by any event.
CI Concentration Increase
This phase follows immediately the anode effect feeding phase. The aim is
to build up the alumina concentration in the bath up to 2 – 2.5 %.
This phase is not interrupted by tapping or anode changing.
When reducing the current load to 50-80% the alumina feed is reduced
correspondingly of 50%.
UA Underfeeding A (normal phase 70%)

This is basically the normal phase and shall lead to a R-slope. The time
period is restricted to 120 minutes. After this time the alumina feed is re-
duced once more (phase B) in order to induce an even more positive R-
slope.
This phase is interrupted during pot operation procedures (anode change,
tapping, manual alumina feed or beam raising).
UA remains active for 60 minutes after the interruption of an anode effect
prediction feed (see OB). The 40% feeding phase follows with the
designation HC: The overfeeding has been stopped because the alumina
concentration is too high. (HC = High Concentration)
UB Underfeeding B (normal phase 40%)

This phase is switched on when the danger of bottom sludging becomes
acute. UB occurs after anode change, after 4 hours since the last anode
effect prediction or at shortside breaking.
This phase is interrupted by an anode effect prediction or an anode effect.
UB is stopped for 30 minutes during tapping. During these interruptions
the respective specific event feeding phases are active in order to prevent
an anode effect.
OA Overfeeding A
This phase is active when the EPC predicts an anode effect. It is a short
intensive feeding which shall prevent the occurrence of an anode effect.
This phase is only interrupted if an anode effect occurs.
OB Overfeeding B
The aim of this feeding phase is to increase the alumina concentration of
the bath.
The alumina concentration increase causes a decrease of the bath resis-
tance. The EPC controls this resistance drop. If 20 minutes after the star
of OA no resistance drop is detected this phase is interrupted: The UA
phase follows for 60 minutes and then the 40% phase UB up to the next

11-10
11. Process Control
anode effect prediction.

Practice has shown that the first anode effect prediction after an anode
change or after shortside breaking is in each case interrupted. This is due
to the increase of the alumina concentration in the bath when changing
anodes or breaking the crust.
After anode setting or shortside breaking the alumina which has dropped
in the bath and eventually arrives under the metal layer has to dissolve.
This phase is switched off during normal pot operation procedures (anode
change, tapping or beam raising).
AC Anode Change
This phase is started by pushing the respective button at the PCT (Pot
Control Terminal).
By reducing the alumina feed after an anode change sludging of the cath-
ode is prevented.
This phase is interrupted when the anode effect prediction is active.
During beam raising or tapping this phase is stopped for the duration of
the event phase.
TP Tapping
This phase is started by pushing the respective button at the PCT.
By increasing the alumina feed the occurrence of an undesired anode ef-
fect is prevented.
CL Current Low
There is an acute danger of an undesired anode effect after a reduction of
the line current especially when the bath temperature went down some
degrees. The EPC cannot detect an eventual coming anode effect. There-
fore a slightly increased feed is permitted for this "blind" time.
BR Beam raising
This phase is started by pushing the respective button at the PCT.
A slightly higher amount of alumina is fed in order to prevent an anode ef-
fect. By pushing the button the bridge is immediately blocked for automatic
movements. Once beam raising is completed, the button is pushed a sec-
ond time in order to reactivate the bridge.
This phase can only be interrupted by an anode effect.
CM Current Middle (reduction to 50-80% of line current)

If the potroom current is reduced to 50 to 80% for more than 5 minutes the
alumina feeding is reduced to 50% of the current phase.
CD Current Drop (less than 50% of the line current)

If the potroom current is reduced by more than 50% the alumina feed is
interrupted. The phase current drop CD is switched on for 30 minutes as
soon as the line current is back to normal (125% alumina feed).

11-11
11. Process Control
11.4 Pot Controller Terminal PCT
11.4.1 General
Figure 11-3 shows the Pot Controller Terminal (PCT).
Figure 11-3: Pot Controller Terminal.

This terminal is desingned according to the wishes of the customer for in-
stance using cyrillic letters for a plant in Russia.
The pot controller terminal provides the following functions for two adja-
cent pots:
• manual entry process information,

• display of pot data,
• call for pot data,
• messages
• enable/disable functions/devices.
The Display:
Figure 11-4 shows the display of the PCT.
Figure 11-4: Display of the Pot Controller Terminal.

The display is two lines, each 40 characters long. The upper line corre-
sponds to the first pot on the pot controller and the second line to the
other. The respective pot number is to be seen as the last item on each

11-12
11. Process Control
line.
The display lines normally show the basic pot measurements. The first
item reading from left to right is the pot voltage. This is followed by a read-
ing of the pot current in kA. The third item from the left is the pot noise in
Pot Noise (μΩ). The third and fourth numbers are the beam positions in
mm. The value before the pot number is the rate in % of theoretical alu-
mina consumption and the accompanying letters give the name of the feed
rate.
When an alarm condition arises the display will flash a text to identify the
alarm. You can stop this flashing by hitting key 9. This will suppress the
alarm message but will leave an illuminated blank display cell just before
the pot number to indicate that an alarm is outstanding. The display will
also help you when signalling pot operations.
The Keys and Buttons:
Select pot1,
Select pot2,
Anode change,
Tapping,
Anode beam raising,
yes,
no,
S functions,
Crust breaking,
F function.
11.4.2 Manual Entry of Process Events
Anode Change:
Shortly before you start to change anode you should signal this action by
pressing the <AC> key for the pot. The sequence is thus:
Press PCU replies “Start with anode change?"

Answer PCU confirms " Changing anode".
The PCU will adjust the voltage setpoint and the feed rate to cope with the
disturbance to the normal pot operation.
At the end press the <AC> key for the pot again.

11-13
11. Process Control
The sequence is thus:
Press PCU replies “Anode change end?"

Answer PCU confirms "Anode change finished".
Tapping:
To start the tapping program hit the <TP> button:
Press PCU replies “Start tapping?"

Answer PCU replies "Tapping xxmm" where xx is the millimetres to
be tapped.
As soon as the pot is ready to be tapped, the tapping light
is switched on.
During tapping the PCU shows the amount to be tapped and the amount
tapped alternately with the normal display of pot voltage and beam posi-
tion etc.
At the end of tapping:

As soon as the computer switches the tapping light off, stop tapping.
If a problem arises during tapping the PCU will signal that it can no longer
follow the tapping operation by flashing the tap lamp and then you should
correct the pot voltage after tapping by lowering the beam yourself.
The computer may not light the tapping lamp immediately, should the pot
be somewaht noisy. It will raise the beam to try to quieten the pot. If tap
start is impossible, the PCU will flash the tap lamp and show the reason
for not tapping:
No tapping - too soon after AE (anode effect),
No tapping - voltage too high,
No tapping - current drop,
No tapping - voltage drop,
No tapping - too little to tap,
No tapping - locked out,
No tapping - beam too low,
No tapping - beam locked,
No tapping - OSP too high,
No tapping - pot too noisy.
Anode Beam Raising:

Just before starting to work on the beam you should hit the button <BR>:
Press PCU replies “Beam raise?",

Answer PCU replies "Beam raise manual".
The green light comes on and you can start the beam raising by hand.

11-14
11. Process Control
After the beam moves have finished, you should refix the beam and:
Press PCU replies “Beam raise finished?",

Answer PCU replies "Beam raise end".
The light goes out at the end of the beam raising operation:
During beam raising the display will show that the operation is taking place
and the normal strict limits on beam moves are temporarily inactive.
Crustbreaking:
Just before staring the work you should hit the button <E>:
Press PCU replies “Start pot servicing?",

Answer PCU replies "Casing or bath service started".
11.4.3 Display of Pot Data

As shown in Figure 11-4 the meaning of the display data is :
4.32: actual pot voltage (V),

106.3: actual potline current (kA),
0.07: calculated noise level (μΩ),
340: bridge position 1 (mm),
341: bridge position 2 (mm),
UA: name of pot feed type,
70%: feed in percentage of theoretical consumption,
: indication of a deviation from the normal pot operation. Acknowl-
edge with key <9>.
>: indicates pot selected,
1104: pot number.
+: indication of a deviation from the normal pot operation, but no

alarm,

11-15
11. Process Control
11.4.4 Call for Pot Data
PCT Information:
You can obtain information on the pot by selecting the pot you want (hit
Pot 1 or Pot 2). The data is seen by pressing one of the number keys:
0: Repeat to show menu list,

1: Anode effect data,
2: Setpoint data,
3: Tap data,
4: Anode change data,
5: Beam raising data,
6: Host data - fluoride and temperature,
7: Host data - bath and metal height,
8: Last overfeed data,
9: Status and alarm display.
S Information:
There are more specialised data available for diagnostic purposes when
using <S> with a number key:
S0: Menu list

S1: Date and time
S2: Nominal current
S3: Overfeed record
S4: Digital input bit readout
S5: Digital output bit readout
S6: Analog input readout
S7: Event bit pattern
S8: Beam status bit pattern
S9: Password entry
11.4.5 Messages
Alarms:
There are a few conditions which the PCU cannot handle or needs to
show to pot tenders. These are flashed on the display line for the pot in
question. By pressing key 9 you suppress the flashing message but this
does not reset the alarm condition. Here is a list of alarms which may
arise:
• Anode beam locked → this is a safety feature, to reset use <F7>,

• Tapping aborted → tap manually, alarm resets itself,
• AE killing failed → kill AE manually, resets itself at end of the AE,
• Output test failed → use <F7> to reset and check outputs.

11-16
11. Process Control
11.4.6 Enable/Disable/Testing of Functions/Devices, Data Input
PCT functions:
There are some actions available by using the F key followed by a number
key:
F0: Repeat 0 to view menu list,

F1: Lock/free resistance control,
F2: Reset noise correction,
F3: Unstarve/starve until anode effect,
F4: Lock/unlock feeders/breakers,
F5: no function,
F6: Lamp test,
F7: Beam lock/unlock and reset,
F8: Feed and break test,
F9: Beam motor test if beam not locked.
Password functions:
There are a couple of functions only available by inputting a password.
They are emergency settings (host down) and diagnostic functions:
P0: Nominal current input - for emergency use (no host communication),
P1: Voltage setpoint input - for emergency use,
P2:
P3:
P4:
P5:
P6: PCT key test,
P7: PCT display cell test,
P8: Send PCT texts to host - (do not use in the potroom).
P9:

11-17
11. Process Control
11.5 Exercises
Exercise 11-1:
Setting of Alumina Feeder
A pot is operated with a current intensity of 115.0 kA at a current efficiency

of 93 %. The pot contains 4.5 tons of bath and consumes 1.9 kilogramm of
alumina for every kilogram of produced aluminum.
The alumina concentration in the electrolyte should be increased from 1.5
to 2.7 wt % with a rate of 0.5 %/h. There are two alumina feeders installed
at the pot. Each feeder adds 1.8 kilgramm per shot to the bath. What is the
feeding frequency and the time period of the overfeeding?
Follow Chapter 5-2 of ElysePrg.
Corresponding to Equation 5-2 of ElysePrg you calculate the alumina con-

sumption in one hour (C(1h)) with:
C (1h) = 0.3356 × 115 × 0.93 × 19

. = 68.20 kg / h
The point feeder setting (fPF) is then according to Equation 5-7 of Ely-
sePrg:
0.5 × 4500
+ 68.20
100 90.70
fPF = = = 25.19 h −1
2 × 18
. 3.6
To determine the time period (tc) to reach the end concentration you use
Equation 5-6 of ElysePrg:
2.7 − 15
.
tc = = 2.4 h
0.5
If you want to use ElysePrg you first introduce the current intensity and the
current efficiency in the Alumina Production page (Page 1) of ElysePrg:

11-18
11. Process Control
and finally on the Point Feeder Setting (Page 3) the other values:
You may also use ElysePrg directly to execute the calculation of Exer-
cise 11-1. Click in the List Examples and Exercises Window of ElysePrg
(Chap. 18.14) on Exs. 11-1 and ElysePrg executes the corresponding cal-
culations.

11-19
11. Process Control

11-20
12. Pot Operation
12. Pot Operation

12.1. Anode Change........................................................ 12-3
12.1.1. Introduction ......................................................................... 12-3
12.1.2. Anode Pattern ..................................................................... 12-3
12.1.3. Anode Changing Cycle ....................................................... 12-4
12.1.4. Anode Changing Procedure ................................................ 12-6
12.1.5. Anodic Current Distribution ................................................. 12-7
12.1.6. Current of Newly Set Anodes............................................ 12-10
12.1.7. Magnetic Effects ............................................................... 12-14
12.1.8. Exercises .......................................................................... 12-15
12.2. Metal Tapping ....................................................... 12-20

12.2.1. General ............................................................................. 12-20
12.2.2. Processing ........................................................................ 12-20

Tel: +41 44 435 3333, Fax: +41 44 435 0665, E-Mail: alesa@alcan
12-1
12. Pot Operation

12-2
12. Pot Operation
12. Pot Operation
12.1. Anode Change
The anode change is divided into six chapters:
• anode pattern,
• anode cycle,
• anode changing procedure,
• anodic current distribution,
• current of newly set anodes,
• magnetic effects.
12.1.2. Anode Pattern

As an example Figure 12-1 shows the anode pattern of an EPT14 pot.
Figure 12-1: Anode pattern of an EPT14 pot.
Fundamentals
In order to eliminate pot disturbances when changing anodes, the se-
quence should be chosen such, that the time period between the change
of two single anodes is as big as possible.
In order to express the quality as a number the "anode space" has been
defined.
This number tells you, how many anode rods you have to jump lengthwise
and in the transverse direction in order to find from the last changed anode
the anode which has to be changed next.

12-3
12. Pot Operation
Here the following principles have to be followed:
• the average value of the 26 anode spaces as big as possible,

• the standard deviation within the 26 anode spaces as small as possi-
ble,
• the change of the corner anodes crosswise and in uniform interval,
• the change with alternating sides,
• the charge of the four hoist gears should be as uniform as possible, in
order to guarantee an anode age in the four sectors as uniform as pos-
sible.
The anode change sequence of EPT-pots has been determined regarding

the above mentioned principles. However, in practice unfortunately some-
times not all these principles can be realized. Therefoer days without an-
ode change should be inserted when small anode spaces arise.
Recently another anode pattern has been introduced in some plants con-
sidering an optimum alumina cover on the anodes to protect these better
against air burn. The anodes are changed side by side as shown in Figure
12-2 for an EPT10-pot.
Figure 12-2: Anode pattern of an EPT14 pot.
12.1.3. Anode Changing Cycle

If anode size, anode weight or current load changes during the operation
of an aluminum electrolysis plant, it is necessary to adjust the anode cycle
to the new conditions.
In order to determine the anode changing cycle the following figures have
to be defined:
• net anode consumption,

• anodic current density,
• apparent anode density,
Net Anode Consumption
There are three different kinds of anode consumption, which have to be
taken into account to get the net anode consumption:
• electrolytic consumption (112 g/kAh see Equation 2-15 for I = 1kAh and
t = 1h),
• non electrolytic excess consumption in the bath,
• non electrolytic excess consumption outside the bath
Generally a separation of the excess consumption is not possible as sepa-

rate measurements cannot be performed (see Figure 12-1).

12-4
12. Pot Operation
For magnetically optimized and microprocessor-controlled electrolysis pots

the non electrolytic excess consumption is about 18 C/kAh. The net anode
consumption (AN(kAh) in kAh) is therefore:
AN(kAh) = 112 + 18 = 130 g / kAh (12-1)
Anodic Current Density

The anodic current density influences directly the anode consumption
I (12-2)
jA =
n A ⋅ l A ⋅w A
Example 12-1: Anodic current density.

You operate a pot with a current load of 140 kA. The pot contains 26 an-
odes of 150 cm length and 52 cm width. What is its anodic current den-
sity?
Apparent Anode Density

The apparent density of anodes depends on the raw materials as well as
on the anode manufacture procedure. The value is taken from the data
sheet, which is supplied with every anode delivery.
Calculation of the Anode Changing Cycle

For the calculation of the anode changing cycle the net anode consump-
tion (AN(kAh)) per kAh is needed. Normally you find however values for the
net anode consumption per kilogram produced aluminum (AN(kg)). For the
determination of AN(kAh) use the following relation:
AN(kAh) = 0.3356 ⋅ AN(kg) ⋅ η (12-3)

You find more information about fractional current efficiency in Chap-

ter 2-3.

12-5
12. Pot Operation
The result is a figure which is specific for a certain plant and is for instance
for an EPT10-pot about 130 g/kAh.
The anode changing cycle is then given with:
hA ⋅S A ⋅ ρ A
t AC = (12-4)
24 ⋅ AN(kAh) ⋅ I
where the usable anode height is calculated with:
h A = h A(meas) − 2 ⋅ σ h
A(meas)
− l Hole − hButt (12-5)
12.1.4. Anode Changing Procedure
Procedure
The anode change is executed with an anode changing vehicle.
This vehicle is equipped with a crane for changing the anodes, a hydrauli-
cally driven wrench to open and fasten the anode clamps on the beam and
a scraper to clean the bath.
Before changing the anode the crust has to be broken. After removing the
old anode the bath is cleaned with the scraper of the vehicle: bigger crust
pieces are removed and the pot has to be skimmed from carbon dust if
necessary. Then the new anode is inserted.
The insertion height is marked on the anode stem.

12-6
12. Pot Operation
In practice two different methods are used to define the insertion height:
• the height taken from the changed anode is transferred to the rod of
the new anode using a hook,
• the insertion height of the new anode is calculated by the computer ac-
cording using Equation 12-6:
H 1 = H 0 − Δh0 + ΔH Br (12-6)
After the change the new anode should be covered with crushed bath and
alumina in order to protect it against air burn. As an insulating layer this
cover is also important for the heat balance of the pot.
Checks after Anode Change

Check Immediately After Anode Change:
The responsible foreman, shift foreman or chief operator examines imme-

diately after the anode change if the pot has METAL WAVES (noisy pot)
checking "DRMM” displayed on the PCU.
Check 12 to 24 Hours After Anode Change:
Another check is made 24 hours after the anode change. The measuring
section controls the current pick up of the last changed anodes. Anode
currents between 50% and 120% of the nominal current indicate a “good”
anode change.
If the current pick up 24 hours after the anode change is less than 50% or
more than 120% of the nominal anode current, the setting range should be
checked and if necessary corrected.
Other varieties in anode setting height are not permitted.
12.1.5. Anodic Current Distribution

The anodic current distribution describes the relationship between the sin-
gle anode currents and the anode age.
Relative Anode Current

Mostly instead of the absolute anode current the relative anode current is
used:
I abs (12-7)
I rel =
I nom
12-7
12. Pot Operation
where the nominal anode current is given with:
I
I nom = 12-8)
nA
Anodic Current Distribution

Figure 12-3 shows a plot of a typical anodic current distribution.
Figure 12-3: Plot of a typical anodic current distribution
One can see that the current of the anodes reaches its nominal value only
after some days. The new anode has been inserted a little higher than cor-
responding to the ACD of the changed anode.
After reaching a maximum the anode current decreases continuously. This
is effected by a decrease of lateral anode area.

12-8
12. Pot Operation
Classification of Current Distribution

The best classification for the quality of a current distribution is the varia-
tion coefficient of the anodic currents. This measurement should be per-
formed shortly before anode change.
σ (12-9)
V A = 100 ⋅
I nom
and for the relative anodic currents (Inom = 1):
Vrel = 100 ⋅ σ rel (12-10)
Table 12-1: Classification of anodic current distribution.
Example 12-2: Anodic current distribution.

For a pot which is operated at 39.7 kA the following anodic currents were
determined:
The standard deviation is 0.85 kA and the variation coefficient conse-

quently VA = 21.6 %.
This is a bad current distribution.

12-9
12. Pot Operation
12.1.6. Current of Newly Set Anodes

New anodes are normally set into the pot with a greater anode to cathode
distance (interpolar distance). The following investigates the influence of
newly set anode on the pot voltage and on the anodic currents.
Consider first a pot with all anodes set to an equal interpolar distance as it
is shown in Figure 12-4.
Figure 12-4: Schematic view of a pot with all anodes set to an equal interpolar distance.
Taking into account the relation for the total resistance of parallel resis-
tances
1
=Â 1 (12-11)
R n RP
and of parallel equal resistances
RP = n ◊R (12-12)
one can write for the proportional resistance under one anode in the elec-
trolyte using Ohm’s Law
UW UW ◊ n A
RBath
0
= = (12-13)
0
I An I ◊ 10 3
12-10
12. Pot Operation
where the anodic current before anode change corresponds to the nominal
anodic current.
The total resistance of one rod i.e. between the anode beam and the liquid
metal is given with
0
RtotAn = RAn
0
+ RBath
0 (12-14)
The entire resistance of all anodes between the anode beam and the liquid
metal is using Equation 12-12
0 (12-15)
R0 = RtotAn
nA
and the corresponding voltage drop
U 0 = 10 3 ◊ I ◊ R 0 (12-16)
When an anode is changed the new anode is inserted higher i.e. the inter-
polar distance of this anode is increased. In addition the contact resistance
stub to anode may be higher which results in a higher resistance between
the anode beam and the anode bottom.
Figure 12-5 shows a newly inserted anode with greater interpolar distance.

12-11
12. Pot Operation
Figure 12-5: Schematic view of a pot with a newly set anode with a greater interpolar dis-
tance.
The overall parallel resistance of the n-1 old anodes (Equation 12-12) is
RPar = R (12-17)
nA - 1
You find the proportional resistance under the new anode by combining
Equation 12-13 with the Equation 4-5 for the electric resistance of the elec-
trolyte with
RBath
new
= RBath
0
◊ IPD + DIPD (12-18)
IPD
The total resistance of the new rod corresponding to Equation 12-14 is

then
new
RtotAn = RAn
new
+ RBath
new
(12-19)
You calculate corresponding to Equation 12-15 the entire resistance be-

tween the anode beam and the liquid metal using Equation 12-11 and
12-19 with

12-12
12. Pot Operation
new
RtotAn ◊ RPar
R1 = (12-20)
RtotAn + RPar
new
and the voltage drop
U 1 = 10 3 ◊ I ◊ R 1 (12-21)
i.e. the pot voltage has changed (Equation 12-16 and 12-21)
DU = U 1 - U 0 (12-22)
You calculate the current load of each anode before anode change using
Equation 12-8 and after anode change of the old anodes with
U1
I 1
An = 0 (12-23)
RtotAn
and of the new (higher inserted) anode with
U1
Inew
An = new (12-24)
RtotAn

12-13
12. Pot Operation
12.1.7. Magnetic Effects

The anodic current distribution changes at the moment when the used an-
ode is removed from the bath.
In the best case the current of the removed anode is distributed uniformly
on the other anodes. In the worst case a single neighbor anode picks up
the current of the removed anode. In practice both cases are not probable.
However, it could be shown that during change of corner anodes the con-
ditions are more complex.
Measurements have shown, that during change of corner anodes the an-
ode of the opposite corner picks up more current than normal. As a result
the current direction changes in the cathode metal. New horizontal cur-
rents are generated changing the metal heave.
Figure 12-6 shows a the local metal upheaval during removal of an anode.
This situation improves however very fast, when the new anode starts to
pick up current.
Figure 12-6: Local metal upheaval during removal of an anode.

It is not appropriate to set an anode higher or deeper due to the
above mentioned reasons; it will act on both corrections with unde-
sired reactions.
An anode which has been set too high will not reach its nominal current
within days; that means that the current has to be picked up from the other
anodes and the anodic current distribution deteriorates.

12-14
12. Pot Operation
12.1.8. Exercises
12-1: Calculation of Anode Cycle.
12-2: Variation Coefficient of Anodic Current Distribution.
12-3: Investigation of Anode Change.

12-15
12. Pot Operation
Exercise 12-1:
Calculation of Anode Cycle:

An EPT14-pot is operated with the following data:
Determine the anode cycle.
Using Equation 12-3 and 12-4 you calculate:

12-16
12. Pot Operation
Exercise 12-2:
Variation Coefficient of Anodic Current Distribution:

The following anodic current distribution (kA) is given:
a) the variation coefficient should be calculated in order to classify the cur-

rent distribution (Anode numbers ≡ anode age).
b) Draw the graph of the current distribution i.e. relative anode currents as
a function of anode age.
Using Equation 10-2 or 14-12 you calculate:

12-17
12. Pot Operation
Exercise 12-3:
Investigation of Anode Change:
An EPT10-pot is operated with the following data before and after anode
change:
a ) calculate the following:
• The anode current(s) before anode change,

• Anode currents of the newly set and of the other anodes
• The increase of pot voltage caused by anode change.
b ) the burning rate of the old anodes is 1.6 cm/day. Calculate the burning
rate of the new set anode for the first 24 hours with the assumption that
the anode currents remain constant during this time.
Using Equation 12-13 and 12-14 you calculate the resistance in the elec-
trolyte under one anode and the total anode resistance with:
Combining Equation 12-15 and 12-16 you write for the total voltage drop:
After anode change you write for the new anode (Equation 12-18 and
12-19):

12-18
12. Pot Operation
With the overall resistance of the other anodes (Equation 12-17) and the
corresponding resistance (Equation 12-20) you calculate the new voltage
(Equation 12-21):
The increase in voltage is (Equation 12-22):
DU = 2.096 - 2.067 = 0.03 V
The anodic current before anode change (Equation 12-8) and after anode
change Equation 12-23 and 12-24 are:
The burning rate of the newly set anode is approximately propotional to

the ratio of anode currents:
This means that the difference of the interpolar distance between the new
anode (which has been inserted 3 cm higher) and the other anodes has
decreased by only
1.6 cm - 1.08 cm = 0.52 cm.
The position of the new set anode is even 24 hours later still about
2.5 cm too high!

12-19
12. Pot Operation
12.2. Metal Tapping
12.2.1. General
The electrolysis pots should always be tapped at about the same time in
the shift. The pots are tapped once every 24 or 48 hours. The tapping pro-
gram is transmitted to the tapping crew in written form. The sequence of
the delivery of the transport crucibles is generally determined by the cast-
house. According to the demand of the casthouse, 3-5 different levels of
metal purity are assembled. It is appropriate to use specially designed
transport and tapping crucibles for metal purities better than 99.9% Al.
12.2.2. Processing
Before tapping the metal level of each pot is measured and fed into the
computer.The computer compares the measured values with the individu-
ally prescribed target value. The difference is equal to the metal height
which has to be tapped.
The metal tapping program is set up by the computer according to the
amount to be tapped and the metal purity which has been announced by
the laboratory for the casthouse product mix.
By pushing the button "TP" the pot PCU is informed that the pot should be
tapped. The computer examines the noise criterion and adapts the alu-
mina- and resistance control respectively.
The signal lamp is switched on when the pot is ready for tapping.
During tapping the PCU controls the resistance continuously and lowers
the beam position correspondingly. As soon as the prescribed tapping
height is reached the PCU switches the signal lamp off. This is the signal
for the operator to stop tapping.

12-20
13. Troubleshooting in
Pot Operation
13.1. Summary .................................................................... 13-3
13.2. Spikes on Anodes ..................................................... 13-4
13.3. Pieces of Carbon in the Electrolyte ......................... 13-5
13.4. Anodes Inserted too Low or Slipped Anodes ......... 13-6
13.5. Anode Set Down on Lumps ...................................... 13-7
13.6. Ridge and Anode Set Down on Top......................... 13-8
13.7. Too Small Interpolar Distance.................................. 13-9
13.8. Metal Roll ................................................................. 13-10

13-1

13-2
13.1. Summary
This chapter treats the identification and correction of problems in

pot operation and the corresponding corrective actions. The follow-
ing problems are described:
Spikes on Anodes,
Pieces of Carbon in the Electrolyte,
Anodes Inserted Too Low or Slipped Anodes,
Anode Set on Lumps,
Strong Side Ledge and Anode Set on Ledge,
Too Small Interpolar Distance,
Metal Roll.

13-3
13.2. Spikes on Anodes

Figure 13-1 shows a spike under an anode.
Figure 13-1: Spike under anode.

A spike on the anode forms whenever one or more of the following condi-
tions is present:
• too much carbon dust in the bath,

• Bath temperature is lower than the liquidus temperature.
Symptoms
The pot voltage tends to drop, consequently the anode beam moves up,
the bath temperature rises, you observe a yellow flame and the affected
anode(s) show high power take-up.
Corrective Actions
• Measure the anodic current distribution in order to identify the affected
anode(s).
• Check the affected anode(s). Replace any anodes with spikes. If sev-
eral anodes have to be changed, use hot anodes with sufficient height
taken from healthy pots.
• Temporary increase of PSP.
• If the aluminum fluoride concentration is high, add soda.
• Skim the pot.
Do NOT change the height at which the anodes are inserted!

13-4
13.3. Pieces of Carbon in the Electrolyte

Figure 13-2 shows carbon pieces in the electrolyte.
Figure 13-2: Pieces of carbon in the electrolyte.

Pieces of carbon in the electrolyte are caused by:
• Poor anode quality,

• Broken rests of anodes remaining from an anode change or from prior
crust breaking.
Symptoms
Increased noise level and varying current in the affected anode(s).
Corrective Actions
• Measure the anodic current distribution in order to identify affected an-
ode(s).
• Check the affected anode(s). Remove the carbon pieces.
• If the pot is still noisy, increase the PSP temporarily.
• If there is a very thin alumina cover on the long side of the pot, break
the crust and free the bottom of the anode with hooks.
• Skim the pot.
Do NOT change the height at which the anodes are inserted!

13-5
13.4. Anodes Inserted too Low or Slipped Anodes

Figure 13-3 shows a slipped anode.
Figure 13-3: Slipped anode.

Cause for anodes inserted too low:
• Insertion mark on the anode rod is at the wrong height
• Incorrect anode setting.
Cause for slipped anodes:

• The clamp was not closed correctly.
• The projecting electrolyte crust on the anode on the other side was not
removed when changing anodes. The anode slips down when the crust
melts off.
Symptoms
High noise level, high current in the affected anode due to the reduced in-
terpolar distance (resulting from the smaller ACD or resulting from the
metal upheaval under the anode due to the higher current), low pot volt-
age, relatively high anode temperature, possibly a red stub.
Corrective Actions
• Pull the anode up IMMEDIATELY!
• Then, ALWAYS, determine the correct insertion height.
It is a mistake to use the anode current measurement as the only

means of determining the correct insertion height!
• Then recheck the anodic current distribution.

13-6
13.5. Anode Set Down on Lumps

Figure 13-4 shows an anode set on a lump.
Figure 13-4: Anode set on a lump.
Cause for anode on lumps:

• Bath not cleaned sufficiently.
Symptoms
High noise level, possible high current in the affected anode, low pot volt-
age, spikes may occur.
Corrective Actions
• Check the anode.
• Clean the bath and then insert the anode again at the correct height.

13-7
13.6. Ridge and Anode Set Down on Top

Figure 13-5 shows a ridge with the anode set on top.
Figure 13-5: Ridge with the anode set on top.

Cause of ridge formation:
• Pot voltage is too low.
• Excess aluminum fluoride concentration is too high.
• The metal level is too high.
Symptoms
High noise level, possible high current in the affected anode and lower pot
voltage.
Corrective Actions
• Increase the voltage.
• Supply additional soda if necessary.
• Reduce the prescribed metal level. Only after the ridge under the an-
odes has been removed: carefully lower the actual metal level.
• Where necessary, set the anode temporarily higher, but then – as soon
as possible – lower it again to the correct insertion height. When doing
so, take into account the reduced burning rate resulting from temporar-
ily raising the insertion height. If necessary, use a hook to transfer the
insertion height from the neighbor anode.

13-8
13.7. Too Small Interpolar Distance

Figure 13-6 shows a too small interpolar distance.
Figure 13-6: Too small interpolar distance
Cause of too small interpolar distance:

• PSP is too low.
• Carbon dust in the bath.
• One or more voltage drops (e.g., bottom voltage) has or have in-
creased.
• Incorrect bath composition having less electrical conductivity
Symptoms
High noise level, the pot shows strong ledge formation under the anode, a
large amount of carbon dust is observable at the tapping and measuring
hole or during anode change, spikes may also form.
Corrective Actions
• Check the bath composition and correct it if necessary.
• Skim the pot.
• Measure the anodic current distribution especially for spikes.

13-9
13.8. Metal Roll
Figure 13-7 shows metal roll.
Figure 13-7: Metal roll.

Cause of metal roll:
Too little density difference between the electrolyte and the metal i.e. the
bath density is too high due to:
• Incorrect composition of the bath, the calcium fluoride content is too
high and/or
• the bath temperature is too low.
Symptoms
High noise level, strong movement within the bath, the aluminum metal
rises, mixes with the electrolyte and moves up to the surface of the bath,
the crust is broken in, unstable anodic current distribution.
Corrective Actions
• Check the composition of the bath and correct it, adding liquid electro-
lyte if necessary.
• Do not do any operation at the pot, changing of anodes, etc. before the
pot has stabilized.

13-10
14. Measurements
14. Measurements
14.1. Summary..................................................................................... 14-3
14.2. Voltage Drops............................................................................. 14-5
14.2.1. Voltage Drop Entry - Anode ..................................................................... 14-5

14.2.2. Bath Voltage ............................................................................................. 14-7
14.2.3. Voltage Drop Metal - Exit ......................................................................... 14-9
14.2.4. Voltage Drop Collector Bars - Exit ......................................................... 14-11
14.2.5. Voltage Drop Metal – Collector Bar........................................................ 14-13
14.2.6. Voltage Drop Entry - Exit........................................................................ 14-15
14.3. Current Distribution Measurements....................................... 14-17
14.3.1. Measurement of the Anodic Current Distribution ................................... 14-17

14.3.2. Measurements of the Cathodic Current Distribution .............................. 14-19
14.4. Weights ..................................................................................... 14-20
14.4.1. Notice ..................................................................................................... 14-20

14.4.2. Metal Content and Current Efficiency by adding Silver.......................... 14-20
14.5. Sampling for the Plant Laboratory ......................................... 14-22
14.5.1. Alumina Concentration Curve ................................................................ 14-22
14.6. Temperature Measurements ................................................... 14-23
14.6.1. Bath Temperature .................................................................................. 14-23
14.7. Measurement of Length and Width ........................................ 14-24
14.7.1. Metal Height ........................................................................................... 14-24

14.7.2. Bath Height............................................................................................. 14-25
14.7.3. Calculation of Interpolar Distance .......................................................... 14-26
14.7.4. Anode Butts ............................................................................................ 14-29
14.8. Measuring Equipment.............................................................. 14-31

14-1
14. Measurements
14.8.1. Measuring Trailer....................................................................................14-31

14.8.2. Tongs Model 78 ......................................................................................14-33
14.8.3. Prod Model 76 ........................................................................................14-35
14.9. Formulas to Evaluate the Measurements...............................14-36

14-2
14. Measurements
14. Measurements
14.1. Summary
Defined measuring points are of utmost importance for the meas-

urements in order to be able to compare the measurement results
throughout the group.
The measurements are classified as follows:

• voltage drops,
• currents,
• weights,
• sampling for the plant laboratory,
• temperatures,
• lengths and widths.
For all measure is described:

• the purpose of the measurement or calculation,
• the place of measurement and frequency,
• the measuring instruments,
• the evaluation and assessment,
• the calibration or control of the measuring instruments.

14-3
14. Measurements
Figure 14-1: Measuring of the voltage drop Entry – Anode.
Figure 14-2: Data Sheet to register the voltage drop Entry – Anode values.
Figure 14-3: Probability Paper.

14-4
14. Measurements
14.2. Voltage Drops
14.2.1. Voltage Drop Entry - Anode
Purpose of Measurement:
The measurement is intended as a check on all anodic voltage drops such
as beam-anode rod (clamp), aluminum–iron weld and stub-anode. The
theoretical voltage drop is a "design" figure. Any deviation from this mini-
mum value indicates the following problems:
• the clamps are not properly tightened or there are dirty contact points
between rod and beam or the contact surfaces are not in order (burnt
spots),
• the aluminum-iron welds are poor,
• the transition resistance stub-anode is not in order.
Measuring Points and Frequency:

Measurements are performed from the 1/3 or 2/3 rising bus bar one meter
above the potroom floor or from the pot voltmeter plus connection up to
three 30 – 50 % burnt-off anodes (8 to 12 pattern days, 1/3 side bus bar
with anodes 1/3 side, 2/3 side bus bar with anodes center and 2/3 side).
The three measurement points on the anodes are located on the anode
surface, 3 cm away on the short side from the collar plate. (Figure 14-1).
For 50/50% current feeding the measurements are performed from the ris-
ing bus bar to the anodes up to the middle of the pot.
All the pots in a potroom or a plant should be measured once a month. If
this voltage drop increases or if the deviation range is large detailed volt-
age drops must be measured.
Measuring Instruments:
Digital voltmeter according to specification in Chapter 14.8.1.
Evaluation:
The measured values should be registered in a data sheet together with
their mean value and standard deviation as is shown in Figure 14-2.
At the same time the measured values should be divided into classes of
20 mV and entered on Probability Paper as shown in Figure 14-3.
If based on this evaluation different populations are found detailed voltage
drops should be measured.

14-5
14. Measurements
Figure 14-4: Measurement of Bath Voltage.
Figure 14-5: Data Sheet to register the Bath Voltage values.

14-6
14. Measurements
14.2.2. Bath Voltage
The measurement is used to check the pot voltage permanent set point.

The bath voltage is measured on three measuring points per pot (Fig-
ure 14-4).
Measurements are performed between the same anodes as for the en-
trance anode and the pot metal. The metal in the pot is virtually an equipo-
tential. The measurement in the metal can, for example, take place as fol-
lows:
The point of a measuring lance is carefully wrapped up with aluminum foil
to a length of approx. 10 cm (if it is not wrapped up the waiting time using
a cold lance is longer) and is pushed through the electrolyte until it touches
the bottom of the pot. It is then immediately pulled back again about 2 cm.
The measured values, the time since the last crust breaking (for side-
worked pots), the set value and pot voltmeter value are entered in the re-
cord sheet. All pots in a potroom or plant should be measured once a
month. This measurement should be performed together with the meas-
urement entry - anode (see 14.3.1).
Evaluation:
The measured values should be registered in a data sheet together with
their mean value and standard deviation as is shown in Figure 14-5.
At the end of a measurement day of or at the end of month the mean
value should be determined with the standard deviation over the pot group
or measured potroom.
In the mean value obtained in this way the following approximate voltage
drops for the anodes are included:
The directives for the calculations see 14.12.

14-7
14. Measurements
Figure 14-6: Measurement of the voltage drop Metal – Exit.
Figure 14-7: Data Sheet to register the voltage drop Metal – Exit values.

14-8
14. Measurements
14.2.3. Voltage Drop Metal - Exit
This measurement is intended to provide a comparison of different bar
sizes
and of different cathode blocks qualities. Also an eventual bottom crust
formation may be detected and the cathode voltage drop increase during
the pot life may be investigated.

The voltage drop Metal – Exit is measured on one measuring points per
pot (Figure 14-6).
The voltage drop is measured from the pot metal to the current exit, the
current exit being at the 2/3 rising bus bar of the following pot one meter
above the potroom floor. The measuring point in the metal is the same one
as described in 14.3.2. For 50/50% current feeding the next rising bus bar
can be taken for the current exit.
This measurement should be performed together with the measurement of
the bath voltage (see 14.3.2) and entry to anode (see 14.3.1).
Evaluation:
The measured values should be registered in a data sheet as is shown in
Figure 14-7.
At the end of a measurement day of or at the end of month the mean
value should be determined with the standard deviation over the pot group
or measured potroom.
The directives for the calculations see 14.12.

14-9
14. Measurements
Figure 14-8: Measurement of the voltage drop Collector Bars – Exit.
Figure 14-9: Data Sheet to register the voltage drop Collector Bars – Exit values.

14-10
14. Measurements
14.2.4. Voltage Drop Collector Bars - Exit
The measurement of this voltage drop is used to determine the voltage
loss in the bus bars.
Measuring Points and Frequency (Figure 14-8):

Measurements are performed from all collector bar ends to the current
exit of the bus bar in question, 1 m above the potroom floor of the bus bar
of the following pot. The bus bar temperature and the average current load
should be determined during the measurement simultaneously with the
voltage drop.
The measurement is performed only once and is valid until modifications
are made of the bus bars or of the pot type or of the respective potroom.
Digital voltmeter and temperature measuring apparatus with thermocou-
ple according to specification in Chapter 14.8.1.
Reading or determination of the average current load during measuring.
Evaluation:
The mean value is determined from the voltage drops and from the tem-
perature measurements. The voltage drop (mean value) is then converted
to an uniform bus bar temperature of 50 °C using the following expression:
1 1
k Al(T )
=
k Al( 20∞C)
d
◊ 1 + a ◊ T - 20 i
κAl(T): electric conductivity of aluminum at the temperature T (m/(Ωmm2),
2
κAl(20°C): electric conductivity of aluminum at 20 °C (m/(Ωmm ),
T: temperature of the bus bar (°C),
α: thermal coefficient of the electric conductivity.
and by using the following values:
m
k Al( 20∞C) = 35 ,
Wmm 2
a = 0.004.
The values are registered in the data sheet as is shown in Figure 14-9.

14-11
14. Measurements
Figure 14-10: Data Sheet to register the voltage drop Metal - Collector Bars.

14-12
14. Measurements
14.2.5. Voltage Drop Metal – Collector Bar
This value is calculated and used to compare bottom blocks and collector
bar sizes with each other.
Calculation:
This value is calculated from the values of Metal - Exit (see 14.2.3) minus
Collector Bar – Exit (see 14.2.4)
Evaluation:
The mean value of the bus bar voltage drop (Collector Bar - Exit) deter-
mined once is subtracted from the calculated mean value of Metal – Exit.
The standard deviation of the mean value Metal – Exit is taken as un-
changed.
The values are registered in the data sheet as is shown in Figure 14-10.

14-13
14. Measurements
Figure 14-11: Measurement of the voltage drop Entry – Exit.
Figure 14-12: Data Sheet to register the voltage drop Entry – Exit values.

14-14
14. Measurements
14.2.6. Voltage Drop Entry - Exit
The measurement is used to control the pot voltmeter.

Measurements are performed from between the 2/3 bus bars, 1 m above
the potroom floor between 2 pots (current entry – current exit). The meas-
urement is made together with the measurements 14.3.1 to 14.3.3. For 50/
50% current feed any two busbars can be taken.
Evaluation:
The pot voltmeter should be read simultaneously with the measurement
and registered in the data sheet as is shown in Figure 14-12.
If the two values deviate more than 0.05 V, the pot voltmeter should be
calibrated. If the deviation is greater than 0.2 V the voltmeter should be
replaced, repaired and recalibrated.

14-15
14. Measurements
Figure 14-13: Measurement of the Anodic Current Distribution.
Figure 14-14: Data Sheet to register the Anodic Current Distribution values.

14-16
14. Measurements
14.3. Current Distribution Measurements
14.3.1. Measurement of the Anodic Current Distribution
The measurement is used to determine the current take-up in the different
anodes also to determine the current take-up of the anodes as function of
the pattern days. Furthermore with the results of the measurements the
setting accuracy of the anodes is checked.

Differentiation is made between the following measurements:
1. Current take-up of newly set anodes with Tongs Model 78/260 (see
14.8.2). All anodes in operation 24 or 48 hours should be measured.
2. Measurement of the current distribution with Tongs Model 78/260 (see
14.8.2). The current take-up in all anodes of a pot should be meas-
ured. The measuring frequency depends on operational difficulties
such as:
• anodes which have slipped off,
• change in the burn-off factor.
3. Measurement with Prod Model 76 (see 14.8.3). This measurement is
much quicker and is used above all for the determination of anodes
which have slipped off, fallen off, anode rods insulated from the
beam, spikes under the anodes and dust formation under the an-
odes. The measuring frequency is fixed by the reduction superinten-
dent.
Current measurment with Tongs Model 78/260 (see 14.8.2).
Current measurment with Prod Model 76 (see 14.8.3).
Evaluation:
The measurement is registered in the data sheet as is shown in Fig-
ure 14-14.
For directions to calculate the variation coefficient see Chapter 14.9.

14-17
14. Measurements
Figure 14-15: Measurement of the Cathodic Current Distribution.
Figure 14-16: Data Sheet to register the Cathodic Current Distribution values.

14-18
14. Measurements
14.3.2. Measurements of the Cathodic Current Distribution
The measurement is used to control the current take-up of the individual
collector bars.
The current take-up of the individual collector bars or collector bars without
current influence the magnetic behavior of the pot. Furthermore crust for-
mation on the bottom can be determined.

All collector bars of a pot are measured. The measurement should be car-
ried out in the case of difficulties in pot operation (at the request of the re-
duction superintendent) .
Current measurment with Tongs Model 78/400 (see 14.8.2).
Evaluation:
The measurement is registered in the data sheet as is shown in Fig-
ure 14-16.
For directions to calculate the variation coefficient see Chapter 14.9.

14-19
14. Measurements
14.4. Weights
14.4.1. Notice
The determination of the following weight values is treated in the

Guidelines for Calculating the Basic Reduction Figures dated May
1986:
• Gross aluminum production,

• Quantity of consumed alumina,
• Weight of consumed anodes,
• Weight of anode butts.
14.4.2. Metal Content and Current Efficiency by adding Silver
The determination of the metal content in the pots takes place in connec-
tion with the determination of the current efficiency for comparison pur-
poses of test groups.
Dermination of Metal Content:

The determination of the metal content should take place both at the be-
ginning and at the end of a test. The length of the test should be at least
90 days. The metal content should be determined in at least 12 pots (pro-
vided the test group is smaller than 12 pots, all the pots) by means of an
addition of silver.
The silver should not be added as pure silver but as an alloy. This alloy
can be ordered from DF Neuhausen. The delivery time is approx. 2 weeks.
The silver/aluminum eutectic is 69.5% Ag, 29.5% Al and 566 °C. This
composition should be requested when ordering the alloy. At the same
time the exact analysis should be requested from DF.
For this alloy composition, the following quantities should be added per pot:
E 10: 700 g ±1 g
E 14: 950 g ±1 g
In the case of sideworked pots the addition should take place after crust-
breaking and before alumina addition. Furthermore, the addition should
not take place in the pot half where an anode will be changed in the next
shift. As the point of time for the first addition, the first crust breaking after
metal tapping should be used. Before adding the silver, a metal sample
should be taken from every test pot.

14-20
14. Measurements
With the addition of the silver alloy a silver concentration of approx. 80 –

100 ppm should be attained. The metal sample should be analyzed in the
own laboratory and at DF Neuhausen. The metal contents at the start of
the test should be determined on the basis of the average silver concen-
trations of three successively taken samples respectively (possibly cor-
rected via Ag concentration before start). For the silver concentration at
the end of the test the concentration of the last test day minus the concen-
tration of the second last day and the silver addition of the second last day
should be used.
Evaluation:
The metal tapping quantities should be corrected with the differences in
the metal content. The determination of the current efficiency should be
determined via the ampere-hours, as before, also the volt-hours etc. Metal
which is used for anode spraying and does not come from the production
of the test group must be subtracted from the tapping quantity. This only
applies for the metal on the anodes.
In the case of failure of one or several test pots the metal content and the
tapping quantity are not taken into account in the evaluation. Additionally
consumed volt-hours for shutdown pots and start-up (provided they are on
the same volt-hour meter) should be determined and corrected accord-
ingly.

14-21
14. Measurements
14.5. Sampling for the Plant Laboratory
14.5.1. Alumina Concentration Curve
The determination is used above all for experimental purposes. With such
a determination, the alumina concentration should be determined as a
function of time and alumina addition.

In the area of the measuring hole a hole of approximately 25 cm diameter
is broken. About one hour before the start of the test is best. The area
around the hole should be splashed with liquid bath in such a way that no
alumina drops into this prepared hole during the test. The bath sample
should be taken, depending on the test in question, by carefully pushing in
the thin bath crust every half or one hour. At least three molds should be
ready for every pot, into which the bath samples are poured. They are
then left to stand for at least half an hour. The samples are then filled into
prepared paper bags and marked with pot and sample number. The center
piece of the sample, approx. 50 % of it, should be taken. The bags should
then be closed carefully and sent to the plant laboratory. The bath tem-
perature should be measured simultaneously with the samples and regis-
tered with the sample number into a data sheet.
The length of the test depends on the kind of test.
Temperature measuring device (see 14.8.1).
Evaluation:
In the plant laboratory, the alumina content is determined with XRD (cali-
brated with the chemical method according to DF instructions 710279). The
samples should be smaller than 100µm ground with a Gulatti mill.
For computer-controlled pots, the analysis values should be registered

with temperatures in the relevant plot. Furthermore the mean value with
standard deviation should be determined over both measuring series,
temperature and alumina concentration.

14-22
14. Measurements
14.6. Temperature Measurements
14.6.1. Bath Temperature
The measurement is used to determine the bath temperature of the indi-
vidual pots.

A hole is broken on the tapping point of the pot in which the bath tempera-
ture is measured with a thermocouple. The immersion depth of the ther-
mocouple into the bath should be 10 cm. In the case of sideworked pots,
the measurement should be carried out from half to one hour after side
breaking.
All pots should be measured every second day.
Marshall thermocouple, Chromel-Alumel.
Digital temperature measuring device (see 14.8.1).
Evaluation:
The measured values should be registered in a data sheet and the mean
value with standard deviation should be formed over the entire potroom or
the pot group.

14-23
14. Measurements
14.7. Measurement of Length and Width
14.7.1. Metal Height
The measurement is used to determine the metal height and the metal
tapping quantity.

In the measuring hole at the height of the rod center of the anode, a hole
is broken and there the metal height is measured. The metal height to-
gether with the bath height (see 14.7.2) is determined.
The measurement is performed daily before tapping.
Metal and bath measuring rod.
Evaluation:
The measured values are registered in a data sheet and the mean value
with standard deviation is determined for every potroom and pot group.
For directions of calculation see 14.9.

14-24
14. Measurements
14.7.2. Bath Height
The measurement is used to determine the cryolite addition program and
the liquid bath tapping and addition program.

The measurement is carried out together with the metal height measure-
ment (see 14.7.1).
Metal and bath measuring rod.
Evaluation:

14-25
14. Measurements
14.7.3. Calculation of Interpolar Distance

In place of measuring the interpolar distance is calculated. These calcu-
lated values are used above all to determine the changes of interpolar dis-
tance for different operating parameters.
You determine the interpolar distance from the potroom mean value of the
bath voltage (see 14.2.2) by rewriting the relation for the ohmic bath volt-
age (Equation 4-6) and putting together all the other components of the
bath voltage like the Reversible Decomposition Voltage, the Bubble Volt-
age Drop, and the Overvoltages into E = 1.65 V. The factor f = 1.2 con-
tains the area ratio bath/anodes and the anodic fanning factor.
UB - E
ACD = k ◊ f ◊ (14-1)
jA
With the meaning of the symbols:
Example 14-1 shows how to determine the interpolar distance for an elec-
trolyte which contains in one case aluminum fluoride and in the other lith-
ium fluoride.
Example 14-1: Calculation of Interpolar Distance
Two pots are operating with several equal parameters: bath voltage =
3.25 V, anodic current density = 0.8 A/cm2, calcium fluoride content (in
the electrolyte) = 4.0 %, aluminum oxide content = 2.0 %, overheat =
5°C. Estimate the interpolar distances when one pot operates with a bath
ratio = 1.19 (no lithium fluoride) and the other with r = 1.38 and a lithium
fluoride content = 2.5 %.
With the relation of Solheim (Equation 3-8) for the liquidus temperature
and of Híveš (Equation 3-9) for the electric conductivty of the electrolyte
you determine by using ElysePrg the following values:
According to Equation 14-1 you calculate the interpolar distance of the

AlF3 – pot with:
3.25 - 165
.
IPDAlF3 = 12
. ¥ 2.305 ¥ = 5.5 cm
0.8

14-26
14. Measurements
and of the LiF – pot with:
3.25 - 165
.
IPDLiF = 12
. ¥ 2.609 ¥ = 6.3 cm
0.8

14-27
14. Measurements
Figure 14-17: Measurement Width, Length and Anode Stem Distance of Anode Butts.

14-28
14. Measurements
14.7.4. Anode Butts
The measurement is used to determine the residual area of the anodes
and the butt height under the stub (distance bottom edge to anode stub).

Figure 14-17 shows the way to measure the length, the width and the butt
height under the stub. The width of the anode butts is calculated with
a1 + a2
w AB = (14-2)
2
With the meaning of the symbols:
Anode supplier and customer arrange the measuring frequency together.
Tape measure and stop or slide gauge.
Evaluation:

14-29
14. Measurements
Figure 14-18:Front View of Measuring Trailer with Digital Voltmeter.
Figure 14-19: Back View of Measuring Trailer with Measuring Lances.
Figure 14-20: Back View of Measuring Trailer with Thermocouples.

14-30
14. Measurements
14.8. Measuring Equipment
14.8.1. Measuring Trailer

Figure 14-18 to 14-20 shows different views of the Measuring Trailer with
digital voltmeter, lances and thermocouples.
Technical Data:
Supply: 12 V lead accumulator 40 Ah with built-in battery

charger for mains connection 220 V, 50 Hz.
Power consumption: from 12 V battery in operation max. 1A.
Measuring ranges: Direct voltage 0…1000 V,

0…199.9 V,
0…19.99 V,
0…1.999 V,
Thermocouple NiCr – Ni 0…1200 °C,

Thermocouple PtRh – Pt 0…1600 °C.
Initial resistance: for voltage measurement ≥ 1 MΩ.
Linearization for thermo-electric volt-

ages installed for: NiCr – Ni and Pt 10 % Rh – Pt.
Cold soldering points-compensation: installed for 0…70 °C.
Working range: 0...50 °C ambient temperature.
Accuracy of measurements: Voltage ± 0.2 % ± 1 digit,

Temperature ± 0.1 % ± 1 digit,
Linear error ± 1 °C.
Platform trailer: Loading bridge 1200 x 800 mm,

Wheels pneumatic tires.
Measuring cable for voltage: Rewind cable reels,

Cable length 10 m.
Field of Application:
The measuring trailer was specially built for performing voltage drop
measurements on reduction cells and for temperature measurements of
the bath. The high direct current fields even near the reduction pot
scarcely influence the accuracy of the measurements of the digital indica-
tor. The level detectors for the temperature measurement, the voltage
measuring points and the necessary tools can be transported on the plat-
form trailer.

14-31
14. Measurements
Figure 14-21: Tong Model 78/260 to measure Anodic Currents.
Figure 14-22: Measuring Tongs Model 78/400 to measure Currents in Collector Bars.

14-32
14. Measurements
14.8.2. Tongs Model 78
Application of Equipment:
The tongs measure relatively easy direct currents as they occur in reduc-
tion pots. By spanning the current-carrying conductor with measuring
tongs the current is directly read off in kiloamperes from the indicator. The
current anodic and cathodic current distributions of a reduction pot can
are determined in this way. The measuring tongs for anode measurements
(Figure 14-21) are equipped with a handle which makes it possible to
measure via a remote control from the sidewall of the pot.
Tongs without remote control but with a larger diameter are used to meas-
ure the current take-up of the collector bars (Figure 14-22).
Measuring Principle:
The measuring principle of the tongs is based on the relation that the line
integral of the magnetic induction along a closed loop is proportional to the
current which is enclosed of that path.
z z
U ◊ dt = k1◊ B ◊ ds = k 2 ◊ I
(14-3)
When the arms of a tong encircle an electric conductor a voltage is in-

duced in it. An integrator determines the voltage-time integral which is
proportional to the intensity of the electric current i.e. the reading at the
exit of the integrator is directly calibrated in kiloamperes.

14-33
14. Measurements
Figure 14-23: Prod Model 76 to measure Anodic Currents.

14-34
14. Measurements
14.8.3. Prod Model 76

Prod Model 76 (Figure 14-23) permits the measurement of anode rod cur-
rent in aluminum electrolysis pot lines in order to find the variation coeffi-
cient. The coefficient describes the difference in current between anodes
under constant pot line current and is defined by Equation 14-12.
Light weight was a primary objective in the design, so that many pots can
be measured in quick succession without fatigue.
We do not recommend absolute current measurement using this voltage
drop method, because temperature variations in the anode rods affect the
conductivity, and also because thermoelectric voltage is included in the
result.
Principle of measurement:
Prod Model 76 uses voltage drop along 20 cm (~8 inches) of the anode
rod. According to Ohm’s Law (Equation 4-4) the voltage drop is propor-
tional to the current to be determined and to the resistance of the alumi-
num rod.
The temperature coefficient for aluminum 99.5 % has been determined as:
1
a= (14-4)
. +T
2218
In order to minimize the error due to thermoelectric voltage Manganin was

chosen as the material for the points. In the case of 100 °C (180 °F) tem-
perature gradient in the Prod Model 76 measuring span the predicted error
is 0.33 mV.
Errors due to poor contact at the anode rod surface are largely ruled out
by the high input resistance of the amplifier.

14-35
14. Measurements
14.9. Formulas to Evaluate the Measurements

In the following some formulas are shown which are useful to evaluate the
measurements:
Mean Value:
n
∑ xi (14-5)
x = i =1
n
Standard Deviation:
n
∑ ex i − x j 2
(14-6)
σ= i =1
n
Standard Deviation for less than 30 values:

n
Â dx i - x i 2
(14-7)
t ◊s = t ◊ i =1
n
Standard Deviation for more than 30 values:

n
Â d xi - x i
2
(14-8)
2 ◊s = 2 ◊ i =1
n

14-36
14. Measurements
Weighted Mean Value:
Xm = Â
ni ◊ xi
(14-9)
Â ni
Simplified Weighted Standard Deviation:
sm = Â ni ◊ s i2 (14-10)
Â ni
Significance test for the difference between the mean values of two
measurement series:
R = x1 - x 2 - t 1 ◊ s 1 + t 2 ◊ s 2
2 2 2 2
(14-11)
n1 n2
R ≥ 0: difference is guaranteed,
R < 0: difference is not guaranteed.
Variation Coefficient:
σ (14-12)
VK = ⋅ 100.0
x

14-37
14. Measurements
Table 14-1: t-Values for standard deviation.
* for more than 30 values generally t = 2.0 is taken.

14-38
15. Autopsies
15. Autopsies
15.1. Goal ........................................................................... 15-3
15.2. Working Instructions for the Autopsy ..................... 15-4
15.3. Abbreviated Standard Program .............................. 15-6

15-1
15. Autopsies

15-2
15. Autopsies
15. Autopsies
15.1. Goal
Whenever a cathode fails unusually early, its behavior prior to the failure
and its condition after is not the same as those of a cathode that ages
normally. The goal of investigating such early failures is to find out what
these differences are, to draw conclusions from them as to the reason why
they occurred, and to decide what corrective measures can be taken to
prevent similar early failures.
To learn the behavior of the cathode that has failed early, all report forms
dealing with the cathode in question must be studied. Under some circum-
stances, an autopsy may be required to establish its condition after the
failure, i.e., a careful cleaning of the pot, accompanied by measurements,
observations, and special sampling. The samples taken must be sent for
investigation to properly equipped laboratories that will then carry out the
analyses needed.
A team of experienced experts should interpret the results of the investiga-

tions on samples, the autopsy, and the analyses, and prepare suggestions
for corrective actions that can prevent similar early failures.
We will deal here only how to execute an autopsy and provide detailed
working instructions for doing so.
It is important to perform all steps and make the associated observations

very carefully during the autopsy. All measurement results of should be
recorded exactly.
All observations are to be documented with a brief description indicating
their position and whenever possible including photographs. The location
of the photos within the pot must be clearly defined.
Test report forms for the autopsy can be found in Section G 4 of the Cath-
ode Handbook. They can be used both for recording the data and as
check-lists to ensure that the data are complete.
There is a special report form for each stage of the autopsy. Positional
sketches are available for recording notable locations (e.g., cracks, ero-
sion locations, etc.). The locations of such points are marked and num-
bered on these sketches. The brief descriptions follow next to the respec-
tive number, with an indication as to whether or not photos are enclosed.

15-3
15. Autopsies
15.2. Working Instructions for the Autopsy

Mark all test samples taken from the cathode and all photos in order to de-
fine clearly their orientation and location within the pot.
• Carefully clean the cathode. Damage as little as possible the cathode blocks.
• Check the carbon surface for eroded spots, cracks, and other damage.
• Check glued joints for wash-outs and penetration of other materials.
• Measure the width of the glued joints.
• Measure the carbon surface.
• Drill cores down to the lining or down to the cathode bars depending on their
location.
• Mark the samples so as to define exactly their position in the pot. Pack them
air-tight and send them for analysis.
Samples are analyzed in the following way :
Carbon ramming paste:
• General condition (standard testing program).

• Compressive strength horizontally and vertically (in the x- and y-directions).
Bonded joints:
• Bending strength (x-direction).

• Joint width.
• Contents in aluminum carbide, electrolyte, and metal.
Blend to the carbon ramming paste:
• Adhesion.
• Metal penetration.
Carbon material from the block in the middle:
• Bending strength vertically (z-direction).

• Compressive strength vertically.
• Cracks.
Drilled sample of calcined carbon material:
• Sodium penetration.
• Degree of graphitization.

15-4
15. Autopsies
The samples are examined in an appropriate laboratory:
• Carefully dig out the carbon ramming paste.

• Check the condition of the end blocks.
• Remove the end blocks.
• Determine where metal has penetrated behind the end blocks and wherever
possible the path of penetration.
• Check the side lining for damage.
• Remove the concrete between the ends of the collector bars.
• Check the ends of the collector bars for separation of the alloy. Determine the
amount of missing material.
• Cut off and remove the ends of the collector bars.
• Wherever possible, dig out the cathode carbon elements individually.
• Measure the cathode carbon elements (above all, the block height over its
entire length). Check for cracks on all sides. Check the upheaval on the top
side and general condition on the bottom side (particularly the bottom of the
bars).
• Check the glued joints and surfaces in contact with the ramming paste for
penetrated metal and aluminum carbide.
• Pack one cathode disc air-tight together with collector bars and send it for
analysis.
Special microscopic investigations on cross-sections of the element study

the quality of the contact surface between the cast iron and the carbon.
The program for the investigations and the laboratory must be determined
in each case:
• Check the firebrick linings for damage and metal penetration.

• Break open the firebrick lining.
• Check the condition of the vapor barrier.
• Lift out the vapor barrier.
• Check the condition of the insulating bricks.
• Remove the insulating bricks
• Check pot walls and floor for eroded spots. Determine any damage to the
steel structure.

15-5
15. Autopsies
15.3. Abbreviated Standard Program

for Investigating Early Failures
• Measure the cooled cathode including edge joints and possibly glued joints
between the carbon elements.
• Drill dry core samples across the damaged location.
• Drill at least two cores for comparison from undamaged locations.
• Send the core samples for laboratory investigation.
The core samples are checked:
• Visually and when necessary under a microscope.

• For changes in ash level from top to bottom.
• For changes in electrical conductivity from top to bottom.
• Estimate the aluminum carbide content by exposing the samples to air.

15-6
16. Emissions from

Aluminum Electrolysis
16.1. General.................................................................... 16-3
16.2. Carbon Monoxide and Dioxide.............................. 16-3

16.2.1. Source ......................................................................... 16-3
16.2.2. Reduction of the Carbon Oxide Emissions................... 16-4
16.3. Fluorine, Dust ......................................................... 16-4

16.3.1. Source of Fluorine Emissions ...................................... 16-4
16.3.2. Source of Dust Emissions............................................ 16-8
16.3.3. Reduction of the Fluorine and Dust Emissions ............ 16-9
16.4. Gas Cleaning Systems ........................................ 16-10
16.5. Fluorocarbon Gases ( CF4 and C2F6 ) ................. 16-12
16.6. Sulphur Dioxide ( SO2 )........................................ 16-12
16.7. Effluents................................................................ 16-13
16.8. Metal Oxides ......................................................... 16-13

16-1
16.9. Spent Pot Lining (SPL)......................................... 16-15

16.9.1. Introduction.................................................................16-15
16.9.2. Composition ...............................................................16-15
16.9.3. Formation Mechanism ................................................16-16
16.9.4. Reactions of Spent Pot Lining with Water ..................16-17
16.9.5. Treatment of SPL .......................................................16-18
16.10. Literature............................................................ 16-20

16-2
16.1. General
Every pot produces hazardous and non hazardous wastes in gaseous, liq-
uid or solid state, which are relevant for the assessments of occupational
health and ecological impact. To get a full picture of the ecological impact
of the aluminum fabrication, it is necessary to take into account also emis-
sions during bauxite mining, alumina refining, electrical power production
and anode fabrication.
The direct generation by the cells can be quantified as follows ( kg / t Al):
• Carbon monoxide and dioxide (CO and CO2) about 1300 kg, depending
on current efficiency.
• Fluorine (gaseous and particulate) about 30 kg, depending on electro-
lyte composition and operating temperature.
• Dust (alumina, carbon, electrolyte) about 30 kg, depending on alumina
and anode quality.
• Sulphur dioxide (SO2) about 20 kg, depending on the sulphur content of
the anodes and anode consumption.
• Organic fluorine compounds (CF4 and C2F6) about 0.2 kg, depending on
anode effect frequency and duration.
• Varying quantities of effluents from the cathode digging pit and an even-
tual wet scrubber, containing dissolved fluorides and cyanides.
• Very small amounts of metal oxides (Fe2O3, SiO2, ZnO...) evaporating
from the liquid electrolyte.
• Spent potlining (carbon and silicates containing fluorides and cyanides)
about 40 kg, depending on pot life.
Since all these compounds have a negative impact on our environment,

they either have to be eliminated, recycled, minimized or converted into
inert materials.
16.2. Carbon Monoxide and Dioxide
16.2.1. Source
The formation of carbon monoxide and dioxide belongs to the basic reduc-
tion reaction of alumina (Lit. 16-1 see also Electrolysis Equation 2-13):
2ηAl2O3 + 3C = 4ηAl + 3(2η-1)CO2 + 6(1-η)CO (16–1)

16-3
As far as carbon is required for the reduction process, the relation carbon
monoxide to carbon dioxide is given by the current efficiency of the proc-
ess. A state of the art cell with 95 % current efficiency produces 1158 kg
CO2 and 82 kg CO per tonne of aluminum. The air burn of the anodes
(about 20 kg / t Al) produces another 70 kg of carbon oxides, their distribu-
tion depends on the reaction temperature.
Since the current efficiency of modern cells is at 95 % compared with older
technology cells with 85 % current efficiency the production of carbon
monoxide could be reduced by about 60 %, however, the respective car-
bon dioxide evolution was increased by about 30 %.
16.2.2. Reduction of the Carbon Oxide Emissions
Today no regulations are limiting the emissions if the concentration does

not exceed a given threshold value. Carbon monoxide is a toxic gas com-
bining reversibly with the oxygen carrying sites of the hemoglobin mole-
cule, preventing the blood from carrying sufficient oxygen. The MAK value
(maximum concentration value at the work place) is set at 30 ppm
(Lit. 16-10).
As already mentioned the carbon monoxide production can be reduced by
an increase of current efficiency, however, the production of the non toxic
carbon dioxide will be increased at the same time.
16.3. Fluorine, Dust
16.3.1. Source of Fluorine Emissions
Fluorine leaves a smelter in form of gases, liquids and solid waste. Its ori-
gin is always the electrolyte used for the electrolysis process.
W. Haupin and H. Kvande presented 1993 (Lit. 16-2) a mathematical
model of the fluorine evolution from Hall - Héroult cells, which updates an
earlier model (Lit. 16-3) . This paper deals with HF and particulate fluo-
rides leaving the cell with the cell gases. However, sources of fluorides
such as SiF4, CF4 and C2F6 are not included in this model. These theoreti-
cal considerations, based on the partial pressure of the different compo-
nents were only checked against measurements on one type of prebake
anode pot.
The classification as gaseous and particulate fluoride in the literature is
arbitrary and does not show the actual situation at the time of evolution.
Gaseous fluorides remain gases at ambient temperatures as HF, SiF4,
CF4 and C2F6. Volatilized electrolyte will become particulate at room tem-
perature. Alumina and carbon dust will adsorb HF and therefore contribute

16-4
to the particulate fluoride. Particulate fluorides may react with the mois-
ture in the air forming HF and measured as gas. These examples show,
that the total fluorine emission is the only meaningful value to talk about.
Volatilization of Electrolyte:
Most of the particulate fluoride evolved from cells results from vaporization
of the electrolyte. The vapor pressure of molten cryolite with additives has
been studied extensively and there is good agreement among recent in-
vestigators (Lit. 16-4, 16-5, 16-6). The major vapor species of electrolyte
components are NaF, NaAlF4 and its dimer.
Entrained Electrolyte:
Liquid electrolyte (Na3AlF6 containing Al2O3, CaF2 and traces of other elec-
trolyte components) entrained in the pot gas freezes and becomes par-
ticulate as the gas cools. It was found (Lit. 16-3) that 5 to 9 % of the par-
ticulate was entrained electrolyte. The entrainment mechanism is ex-
plained as follows (Lit. 16-7):
When a gas bubble reaches the electrolyte surface it breaks producing tiny
droplets that will be carried away if the gas velocity is sufficient. It has
been shown (Lit. 16-7), that entrainment varies directly with the gas veloc-
ity and inversely with the surface tension of the electrolyte. In this case
only the surface tension varies significantly. The electrolyte crust acts di-
rectly as a filter to remove entrained electrolyte and a long exit path gives
the droplets more time to settle out.
Hydrogen Fluoride:
Hydrogen fluoride (HF) is produced by hydrolysis of electrolyte or hydroly-
sis of electrolyte vapor. It is assumed, that moisture in the alumina reacts
with aluminum fluoride according the formula:
2AlF3 + 3H2O = 6HF + Al2O3 (16–2)
Since the equilibrium constant for the other electrolyte components is at

least a factor of 1000 higher it is unlikely that they react in significant quan-
tities with the moisture.
Another source of hydrogen fluoride is the hydrolysis of electrolyte vapor
with moisture in the air according the formula:
3NaAlF4 + 3H2O = 6HF + Al2O3 + Na3AlF6 (16–3)
The incoming air supplies the moisture and on the other hand cools the
vapor, limiting hydrolysis. While the hydrolysis of electrolyte vapor
changes the amount of gaseous fluoride, it does not change the total fluo-
ride emission.
The predictions of this mathematical model are shown in the following fig-
ures (Lit. 16-2).

16-5
Figure 16-1: Evolution of gaseous and particulate fluorine in dependence of the bath ra-
tio. The overheat (ΔT) was kept constant.
Figure 16-1 shows the evolution of gaseous and particulate fluorine in
function of the aluminum fluoride content of the electrolyte at different
temperatures, Figure 16-2 the increase of evolution with decreasing alu-
minum oxide content in the electrolyte.
Figure 16-2: Increase of evolution with decreasing alumina content in the electrolyte.
Figure 16-3 shows the rather dramatic decrease of evolution produced by

adding 3 % lithium fluoride to the bath at constant 15 °C superheat.

16-6
Figure 16-3: Influence of lithium fluoride additions on fluoride evolution.

Figure 16-4 predicts the entrained bath evolved from a modern cell. It is a
minor factor and mainly a function of alumina concentration.
Figure 16-4: Entrained bath evolved from a modern cell.

For all predictions the barometric pressure, the absolute humidity, H2O in
the alumina, H in the anodes, fluorspar and magnesium fluoride in the
electrolyte were held constant.
These calculations show a decrease of the fluoride evolution with increas-
ing bath ratio, increasing alumina and lithium fluoride concentrations in the
electrolyte and decreasing superheat. The entrained electrolyte at 15 °C
superheat has its maximum at about 4 % alumina concentration in the

16-7
electrolyte. Decreasing aluminum fluoride concentration will increase the

amount of entrained electrolyte.
Comparing these curves with values measured in the Alusuisse smelters
Essen (Lit. 16-8) and Steg (Lit. 16-9) we see that the 31 kg F / t Al meas-
ured at Essen fit well into the predicted range. Steg’s value of 51 kg F / t Al
suggests, that the superheat at the time of measurement must have been
above 25 °C. Unfortunately electrolyte temperature and composition with
the exception of aluminum fluoride were not measured and the probable
measuring error estimated by the authors is 15 %.
16.3.2. Source of Dust Emissions
Dust emitted from the cells comprises > 70 % of alumina, carbon dust,
condensed volatiles from the electrolyte and entrained electrolyte. The
quantities mentioned in the literature vary from 20 to 50 kg / t Al (Lit. 16-8,
16-11).
The analysis of the particle size distribution of cell duct gas (Lit.16-12)
shows two peaks (Figure 16-5), one at about 0.2 mm and the other at
about 20 mm. Along with condensed volatiles, the fine fraction is also
strongly enriched in impurities, such as Fe, Si, P, S, Ca, V, Ni.... Alumina
containing a high proportion of crystalline water or absorbed volatiles can
undergo breakdown during feeding, generating fine dust. The particle
breakdown can also occur during dry scrubbing, especially if multiple
passes are used.
Figure 16-5: Typical particle size distribution of dust evolved from primary aluminum cell.

16-8
16.3.3. Reduction of the Fluorine and Dust Emissions
In most of the countries the emissions of fluorine and dust are more or
less stringent limited by legislation, however, not all these laws are strictly
applied. The standard in Europe allows usually a specific emission of 0.8
to 1.0 kg Ftot / t Al for existing smelters. EPA is now proposing a standard
of 0.6 kg Ftot / t Al for new smelters. Very often the quantity of fluorine re-
leased per year is also limited, as for example for the smelter at Steg.
Normally the HF concentration at the stack of a dry scrubber should be < 1
mg HF / m3. In various countries not all types of smelters have to comply
with these most stringent limits. Exemptions for old technologies are
sometimes applicable.
The MAK values given are as follows (Lit.16-10):

HF: 3 ppm
Ftot: 2.5 mg / m3
Dust: 2.5 mg / m3
The cell operation has very limited influence on the fluorine and dust evo-
lution (see 16.3.1) of the cells. New state-of-the-art cells are already de-
signed for optimum operation with high current efficiency. The operation
has to achieve stable conditions in the cells at a superheat of < 10 °C and
few anode effects, which gives the lowest fluoride emission. The raw ma-
terials like alumina and anodes mainly contribute to the dust evolution. For
a low dust emission the alumina used must have a low α - alumina content
(< 10 %), << 10 % fine particles (< 45 μm) and a small attrition index.
On the other hand the working discipline of the operating personnel has a
major influence on the collection efficiency of the pot gas. Early opening of
hoods prior to anode change or tapping as well as substandard mainte-
nance of the collection system may reduce its efficiency far below 90 %.
Inefficient maintenance and / or insufficient spare capacity of the cleaning
system have a very detrimental effect on the cleaning efficiency. These
factors decide whether a smelter complies with the given regulations or
not.
The modern aluminum industry can be considered as a front runner in en-
vironmental control. A state-of-the-art smelter with hooded cells and a dry
scrubbing system releases << 1.0 kg Ftot / t Al and about 1 kg dust / t Al.
This performance can only be achieved with a collection or hooding effi-
ciency of > 98 % and >99.5 % cleaning efficiency in the dry scrubbing
system, as shown on Table 16-1.

16-9
Table 16-1: Ftot Emissions (kg F / t Al). Total evolution from the cell: 35 kg Ftot / t Al.
Collection efficiency 90 % 95 % 97 % 98 % 99 %
Scrubbing efficiency 98.5 % 4.0 2.2 1.6 1.2 0.9
Scrubbing efficiency 99.7 % 3.6 1.8 1.2 .8 0.5
16.4. Gas Cleaning Systems
State-of-the-art gas cleaning systems require a highly efficient hooding

suction system to capture the gas evolved by the pots. To capture the few
cubic meters of reaction gas between 4000 and 6000 m3 of air / gas mix-
tures are evacuated from the cells and transported into the scrubbing sys-
tem.
Wet Scrubbing:
In wet scrubbing systems, the collected gas is cleaned with neutral or
slightly caustic water. Nowadays wet scrubbers are outdated, because the
emission problems are only shifted from the air into the water and reclaim-
ing the valuable material from the water is tricky and cost intensive.
Dry Scrubbing:
The dry scrubbing process developed about 35 years ago allows the easy
recycling of valuables (alumina and fluorides) in the pot gas. The re-
claimed fluorides and alumina usually pay only for the operation costs of
the scrubber, the financial costs have to be booked against environmental
protection.
A dry scrubbing system comprises:
• Pot exhaust ducting system up to the bag filter system

• Reactor
• Bag filter
• Alumina handling
• Operation and emission control systems
Its requirements are:

16-10
• Even and constant gas exhaust from the cells.

• High cleaning efficiency of HF and dust.
• Moderate gas velocities in the reactor and filter system in order to pre-
vent alumina attrition, reduce scaling, minimize wear and pick up of im-
purities and increase the reaction time.
• Flexible system with on stream and environmentally acceptable mainte-
nance condition.
• Stand by units for vital equipment.
There are principally two different types of reactors in operation, one

based on the principle of a vertical Venturi pipe, where the alumina is in-
jected at the bottom and flows together with the gas into the bag house.
The reaction time available is very short and a significant portion of the HF
is finally adsorbed on the filter cake in the bag house. The second is a fluid
bed type reactor, where the alumina is injected crosswise into the gas
flow with a longer reaction time between HF and alumina.
The alumina required for a good cleaning efficiency should have a specific
surface of > 50 m2 / g Al2O3. For the cell operation itself an even higher
specific surface is asked for. This alumina giving good operation results
with state-of-the-art cells is also favorable for the dry scrubbing of the pot
gas.
HF Adsorption on Alumina:
The reaction
Al2O3 + 6HF = 2 AlF3 + 3H2O (16–4)
has unfavorable kinetics and goes very slow at the prevailing tempera-
tures.
The favorable reaction is a chemisorption of the HF - molecules on the
alumina. That means that the HF has a chemical bonding with the OH
groups of the γ - alumina. There will only be a single layer of HF molecules
absorbed on the surface of the alumina. The chemisorption rate increases
with temperature. According to Cochran (Lit. 16-18) the percentage of HF
that can be chemisorbed onto alumina can be calculated by dividing the
surface area of the alumina by 33. An alumina with a BET surface of 60 m2
is saturated with 1.8 % HF.
A physical adsorption based on inter-molecular bonding forces in multi
layers is also possible, however, the bonding energy is low and decreases
with temperature. The physically adsorbed HF re-evolves upon heating
and adds a recycle load on the recycle system.

16-11
16.5. Fluorocarbon Gases ( CF4 and C2F6 )
Source:
CF4 and C2F6 are considered as non toxic greenhouse gases, which may
contribute to the global warming. The presence of CF4 in anode gas during
anode effects is well known since many years. During normal electrolysis
the potential is too low for the CF4 formation, since the standard potential
of the reaction
4Na3AlF6 + 3C = 4Al + 12NaF + 3CF4 (16–5)
is E0 = 2.50 V compared with 1.16 V for the reaction of CO2 formation from
alumina and carbon.
Henry and Holliday (Lit. 16-13, 16-14) reported about 40 years ago, that
the gas from industrial cells contains about 14 % CF4 and 0.1 % C2F6 dur-
ing an anode effect. More recent measurements from ALCOA showed 1.2
kg CF4 / t Al and 0.02 kg C2F6 / t Al (Lit. 16-15). A survey in Canadian
smelters gave 0.5 to 1.0 kg CF4 and 0.02 to 0.08 kg C2F6 per anode effect
(Lit. 16-16). 1996 measue data from AP 18 and AP 30 cells (Lit. 16-15)
showed that the overvoltage influences the formation of CF4 and C2F6.
Reduction:
Since the fluorocarbon gases cannot be trapped in normal dry or wet
scrubbers the only way to reduce the emission is to avoid anode effects.
State-of-the-art point feeder pots with prebaked anodes and good process
control allow an operation with about one anode effect per pot and month,
compared with about 2 per pot and day 40 years ago. That means that the
CF4 emission of about 1.6 kg / t Al could be reduced to about 0.02 kg / t Al
just by reducing the number of anode effects.
16.6. Sulphur Dioxide ( SO2 )
Source:
SO2 is a toxic gas and contributes to the acid rain if released into the air.
The MAK value is set at 2 ppm, which is practically never exceeded in the
potrooms. The main source of SO2 emission is the sulphur content in the
anodes, however, the contribution of the alumina should not be neglected.
Depending on the type of fuel used for the calcination of the alumina, its
sulphur content may be as high as 0.1 %, causing SO2 emissions up to
about 4 kg / t Al. During electrolysis the oxygen discharged at the anode
combines with the sulphur in the anodes forming SO2. This reaction con-

16-12
tributes about 20 kg SO2 / t Al if the sulphur content of the anodes is about

2 %.
Legislation for limiting the SO2 emissions is not yet as far as for fluorides.
Specific emissions of up to 4 kg SO2 / t Al are usually tolerated. Limitations
of the total SO2 emission per year may force new smelters to install spe-
cial scrubbing units.
Reduction:
Long times it was possible to meet the emission limitations by limiting the
sulphur content in the anodes to 1.5 % to 2 %. Since the quantities of low
sulphur coke for anode fabrication are limited, we either have to pay a
premium or have to install SO2 scrubbers.
The standard SO2 scrubber is installed after the dry scrubber and is oper-
ated with sea water, if available. The SO2 is dissolved in sea water and
discharged into the sea. Inland fresh water, containing soda, has to be
used, cleaned and recycled.
16.7. Effluents
Besides the sanitary effluents and cooling water from rectifiers, eventual
wet scrubbers and casting pits a smelter has to take care of different small
sources which very often contain toxic components. The most critical are
eventual effluents from the cathode relining shop. If the spent cathodes
are cooled with water or dug out outside a building all the water leaching
from the cathodes has to be collected and treated. The same applies for
water leaking from open-air storage areas for spent pot linings. All these
effluents are caustic, containing sodium fluoride (NaF) and sodium cyanide
(NaCN) leached from the spent pot lining. These effluents have to be oxi-
dized to destroy the cyanides, neutralized and normally filtered before
leaving the premisses.
16.8. Metal Oxides
Metal oxides (Fe2O3, SiO2, Ga2O3, ZnO....) dissolved in the electrolyte

have also the tendency to evaporate and will be recycled via the dry
scrubbing unit, which means that the resulting metal purity from a hooded
pot will be worse than metal from an open pot. Consequently a small por-
tion of the metal oxides will be released into the air.
The distribution coefficients (k) gas / liquid aluminum for various metals
were already reported some 20 years ago (Lit. 16-19 to 16-22) the most
important are listed in Table 16-2:

16-13
Table 16-2: Distribution Coefficient (k) Gas / Aluminum. % of impurity which goes
into the metal, open cells.
Si Zn Ni Fe Ga Ti V P Pb
k 0.1 0.3 0.7 0.8 1.0 1.5 2.8 6.0 2.0
% into metal 95 80 60 55 50 40 25 15 33
For the calculation of the resulting impurity level in the aluminum the rela-
tioship (Equation 16-6) will be used.
S (16–6)
M=
a f
1+ 1− η ⋅ k
The difference between S and M escapes into the air.
Examples for open cells:
Examples for hooded cells, collection efficiency 98 %:

16-14
16.9. Spent Pot Lining (SPL)
Today the spent pot lining from aluminum electrolysis cells is a major envi-
ronmental concern of the aluminum industry. On the other hand SPL has a
good recovery potential because of its fluoride and energy content. Some
dozens of patents cover the recycling of SPL as for example recovering of
the fluorides, fabrication of ceramic tiles, as fluoride supplier in the steel
industry, as fuel in cement kilns, recycling of the carbonaceous material in
ramming paste or anodes, reuse of crushed insulation material in the cell
relining process.
Since SPL is generally classified as hazardous waste the handling and
disposal of SPL is highly regulated by legislation. However, the require-
ments vary from country to county and are often updated, changed and
challenged.
16.9.2. Composition
SPL is the remaining of a cathode at the end of its useful life of several
years. The quantity depends on the size and design of the cell and most
important of lifetime and amounts to about 40 kg SPL / t Al.
SPL comprises the carbon cathode, refractory materials, penetrated alu-
minum and electrolyte components, with minor amounts of aluminum car-
bide and nitride as well as cyanides. According different references (Lit.
16-23, 16-24, 16-25, 16-26), SPL contains about 30 % carbon, 30 % re-
fractory, up to 40 % fluorides and about 0.2 % cyanides, if the collector
bars are not counted.
The exact composition of the SPL varies considerably from plant to plant
and from cell to cell. The composition depends on lining design, choice of
materials, cell life, degree of aluminum and electrolyte penetration.
Table 16-3 shows an analysis given by Bings (Lit. 16-24), Table 16-4 one
given by Radvila (Lit. 16-25).
Table 16-3: Analysis of Spent Pot Lining (Lit. 16-24).

Mass %
Graphite 30
Al2O3 26
Cryolite 13
NaF 8
CaF2 3
Quartz 3
Na2CO3 6
Other 11
Table 16-4: Analysis of a 100 kA cell after 3½ years operation (Lit. 16-25).

16-15
Bottom Sidewall Average

% Na 15.5 19.0 16.7
%F 11.8 14.7 12.8
% Al* 10.5 11.4 10.7
* incl. Almet 0.1 0.1 0.1
* incl. Al4C3 0.3 1.4 0.7
* incl AlN 2.7 1.6 2.3
Al2O3 4.5 5.8 5.0
SiO2 11.2 10.0 10.8
C 37.2 29.0 34.4
Ca 1.3 2.5 1.7
Fe 1.3 0.8 1.1
SO4-- 0.8 0.8 0.8
CO3-- 1.4 1.5 1.4
CN- 0.03 0.4 0.2
Sodium cyanide as well as sodium ferrous / ferric cyanide and the fluoride
content of SPL are the main environmental concerns. All cyanides and
most of the fluorides are water soluble and leachate from SPL inventories
may contaminate groundwater or effluent from the smelter. SPL should be
stored under a shelter in a dry area! Leachates from untreated SPL dis-
posal sites have highly variable compositions (Lit. 16-27). An analysis of
leachate from spent pot lining disposal sites is shown in Table 16-5.
Table 16-5: Analysis of Leachate from Spent Potlining Disposal Sites.
Range
pH 10 - 12
Total Dissolved Solids 20 - 30 g / l
Total Suspended Solids 10 - 3000 mg / l
Total Organic Carbon 300 - 500 mg / l
Carbonates 6 - 15 g / l
Sulfates 0.8 - 7 g / l
Fluoride 1.2 - 8.5 g / l
Total Cyanides 0.15 - 1.25 g / l
Total Sodium 6.5 - 15 g / l
16.9.3. Formation Mechanism
The initial reaction leading to the impregnation of cathodes by salt is:
Al(liquid) + 3NaF(Cryolite) = 3Na(Carbon) + AlF3(Cryolite) (16–7)
This reaction will increase with decreasing aluminum fluoride excess in the
electrolyte.
If air is allowed to penetrate (bad sealed collector bars!) the carbon lining
the following four reactions may proceed:
4Na3AlF6 + 12 Na + 3O2 = 2Al2O3 + 24 NaF

16-16
(16–8)
2Al + N2 = 2AlN
2Na + 2C + N2 = 2NaCN
4Na + 3O2 = 2Na2CO3

The cyanide distribution in a cathode shows a characteristic pattern. The
highest are in the sidewall (Table 16-4) and are decreasing in the direction
to the center of the cell were less air respective nitrogen is available. The
cyanide formation from the elements already starts at about 530 °C
(Lit. 16-28).
Metallic aluminum may be formed by the reaction according to Equa-
tion 16-9. This reaction will work in both directions depending on the reac-
tion temperature. The formation of sodium is favorized at elevated tem-
peratures.
Na3AlF6 + 3Na = 6NaF + Al (16–9)
16.9.4. Reactions of Spent Pot Lining with Water
If SPL gets into contact with water reaction gas will evolve containing hy-
drogen (H2), methane (CH4) and ammonia (NH3), leaving sodium hydrox-
ide, sodium aluminate and alumina. The main reactions are:
2Na + 2H20 = H2 + 2NaOH
2Al + 3H2O = H2 + Al2O3

(16–10)
2AlN + 3H2O = 2NH3 + Al2O3
Al4C3+6H2O = 3CH4 + 2Al2O3
2Al + NaOH + H2O = 3H2 + 2NaAlO2

Mixed with air, hydrogen and methane may form an explosive gas and it is
not recommended to keep SPL in closed rooms with humid air.

16-17
16.9.5. Treatment of SPL
Several technically feasible alternatives exist or have been proposed for

the treatment of SPL.
Fuel Supplement in Power Stations:

The carbonaceous material has a heat value of about 4800 kcal / kg,
which corresponds to carbon with 35 % ashes. It was tested in a thermal
power station (Lit. 16-29). 2 % cathode carbon, separated from SPL, was
added to the fuel. An analysis showed no cyanides at the stack, however,
the fluoride emission with the waste gas increased from 3 mg / m3 to
11 mg / m3.
Fuel and Raw Material in the Cement Industry:

Since the combustion air in cement kilns is preheated to very high tem-
perature by heat exchange with hot clinker, thus making it possible to burn
almost any kind of fuel. A replacement of coal is possible if the kiln has a
burner for pulverized coal. Due to the high fluorine content (10 - 15 %) the
quantity is limited (Lit. 16-30) to about 0.8 % of the clinker. The high tem-
perature of the flame (1800 - 2000 °C) in the kiln assures that the cyanides
will be converted into CO2 and NOX.
The benefits and the problems associated with the use of SPL can be
summarized (Lit. 16-31, 16-32) as follows:
• SPL can save fuel cost

• Additional fluoride emissions from the kiln are negligible
• Complete cyanide destruction
• Fluoride in SPL allows a lower calcining temperature
• Cement chemistry and quality within specifications
• The high sodium content of SPL makes it unsuitable for low alkali ce-
ments
• Excess alkali and fluoride may cause kiln operational difficulties
• SPL effect on long term cement durability is not known.
Slag Additives in the Iron Industry:

SPL together with limestone additions gave a good slag in a pilot plant cu-
pola melting gray iron (Lit. 16-32). In this respect SPL is a good replace-
ment for fluorspar.
Leaching Process:
The main problem of all leaching processes is how to dispose off the cya-
nides. Most of the cyanides in the leachate are complexed with iron and
are difficult to oxidize chemically. The recovery of the fluorides is either in
form of fluorspar or cryolite. Unfortunately modern smelters have an ex-
cess of cryolite and it is nowadays very difficult to sell this recovered cryo-
lite.

16-18
Combustion, Calcining:
1993 Reynolds Metals started the commercial operation of its SPL treat-
ment facility in Arkansas (USA). This plant is designed to convert SPL into
an environmentally acceptable material. The Reynolds process flow sheet
(Figure 16-6) combines the most important elements of successful SPL
treatment: competitive cost, complete health protection for everybody,
complete recycling of all raw materials, elimination of long term liabilities to
the SPL producer.
Figure 16-6: Flowsheet of the Renolds Metals Inertization Process.

In this process the SPL is mixed with limestone and an anti-agglomeration
agent (sand) and burnt in a rotary kiln. The residual product has been de-
listed by the US Environmental Protection Agency and it is expected to sell
it in the near future as raw material for asphalt aggregate, cement feed-
stock, aggregate base and concrete shapes.
At the Boyne Island smelter of Comalco the so called “Comptor” process is
now under a full scale test with a throughput of about 10000 tpy SPL. It is
a relatively simple 3 stage process. The first stage is a crusher to produce
SPL with particle size < 0.6 mm. The second stage is a calciner heating
SPL to about 700 °C. In the presence of water and oxygen the cyanides
will be broken into nitrogen and CO2. The leachable cyanide content of the
calcined product is about 200 times lower than the limits set by the authori-
ties. In the last stage, the fluorides were recovered in form of aluminum
fluoride. After removal of the fluorides, the residue passes a standard
leach test and is no longer considered toxic.
More details of other processes can be found in Aluminum (Lit. 16-33 to

16-36).

16-19
16.10. Literature
Lit. 16-1: T.G. Pearson, J. Waddington,

"Electrode Reaction in the Aluminium Reduction Cell",
Discuss.Faraday Soc. (1947), 1, 307-320.
Lit. 16-2: W. Haupin, H. Kvande,

"Mathematical Model of Fluoride Evolution from Hall-Héroult Cells",
Lit. 16-3: W. Wahnsiedler, R. Danchik, W. Haupin, D. Backenstose, J. Colpitts.

"Factors Affecting Fluoride Evolution from Hall-Héroult Smelting Cells"
Light Metals 1978, V 2, 407
Lit. 16-4: J. Guzman.

Dr. Ing. Thesis University Trondheim 1986.
Lit. 16-5: H. Zhou, O. Herstad, T. Østvold,

"Vapour Pressure Studies of and Complex Formation in NaF-AlF3 and
Na3AlF6-MgF2 Melts",
Lit. 16-6: H. Kvande.

Dr. Techn. Thesis University of Trondheim 1979.
Lit. 16-7: J. Miller,

"Environmental Conditions Inside and Outside a Modern Aluminum
Smelter at A/S Årdal og Sunndal Verk (ASV) Norway",
Proc: Engr. Aspects of Pollution control, Metals Soc. London (1974),
165-172.
Lit. 16-8: M. Frisch, A. Sigg,

A-L Umweltservice Bericht 90-34, 16. 08. 1990.
Lit. 16-9: M. Frisch, A. Sigg,

A-L Umweltservice Bericht 90-39, 23. 08. 1990.
Lit. 16-10: IPAI,

The measurement of employee exposures in aluminium reduction plants,
2nd Edition Sept. 1991.
Lit. 16-11: J. B. Metson, B. J. Welch, M. M. Hyland and J. J. J. Chen,

"Gaseous Emissions from Aluminium Reduction Cells",
Chemistry in New Zealand, Jan. 1995, 19-24.
Lit. 16-12: K. Grjotheim, B. J. Welch,

"Aluminum Smelter Technology",
2nd Edition,
ISBN 3-37017-162-6.
Lit. 16-13: J. L. Henry, R. D. Holliday,

J. Metals 9, 1957, 1384.

16-20
Lit. 16-14: R. D. Holliday, J. L. Henry,

Ind. Eng. Chem. 51, 1959, 1289.
Lit. 16-15: R. A. Roberts, P. Ramsey,

"Evaluation of Fluororcarbon Emissions from the Aluminum Smelting
Process",
Lit. 16-16: H. I. Schiff, J. Bechara, J. T. Pisano, G. I. Mackey,

"Measurements of CF4 and C2F6 in the Emissions from Alumium Smelt-
ers",
Proc. SPIE - Int. Soc. Opt. Eng. 1994, 81.
Lit. 16-17: G. Bouzat, J. C. Carraz, M. Meyer,

"Measurements of CF4 and C2F6 Emissions from Prebaked Pots",
Lit. 16-18: C. N. Cochran, W. C. Sleepy, W. B. Frank,

"Fumes in Aluminum Smelting: Chemistry of Evolutuonand Recovery",
J. Metals, 22 (1970) 9, 54-57.
Lit. 16-19: V. Sparwald,

"Volatilization of Impurity Elements during Electrolytic Production of
Aluminium",
Erzmetal 26 (1973) 11, 529-533.
Lit. 16-20: V. Sparwald,

"The Source of Impurities in Primary Aluminum",
Aluminium 54 (1978) 10, 629-631.
Lit. 16-21: R. E. Frankenfeldt, U. Mannweiler,

AIME 1975.
Lit. 16-22: K. Grjotheim, K. Matiasovsky,

Aluminium 1983, 687.
Lit. 16-23: W. S. Rickman,

"Circulating Bed Combustion of Spent Potliner",
Lit. 16-24: N. H. Bings, F. W. Kämpf, H. Sauer, H. W. Schmidt, J. N. Andersen and J. B.

Todd,
"Pyrohydrolysis of Wastes from Electrolytic Aluminum Reduction Sys-
tems",
Erzmetal 37 (1984), 435-441.
Lit. 16-25: P. Radvila, J. Bruppacher, H. Hadorn,

"Charakterisierung und Aufarbeitung von Kathodenausbruch",
Report 7608/78 A-L R&D Neuhausen (1978).
Lit. 16-26: J. E. Deutschman, J.S. Lobos, D. O. Johnson, A. S. Reid,

"A Process to Produce from Potlining in an Inert Residue Suitable for
Landfill Sites",

16-21
Lit. 16-27: R. Keller, K. T. Larimer,

"Experimental Study on the Reaction Forming Cyanide in Potlining",
Lit. 16-28: H. L. Givens,

"Using Spent Potliner as a Fuel Supplement in Coal-Fired Power
Plants",
JOM, 41 (1989) 3, 57-59.
Lit. 16-29: J. A. Imlach,

"Disposal of Spent Potlining: Research and Development",
Report MA88/3567/D 26.7.1988, p. 1023.1030. Alusuisse Erze GmbH.
Lit. 16-30: R. L. Byers,

Light Metals 1982, 1023.
Lit. 16-31: R. C. Dickie,

"Spent Potlining as a Fuel Supplement in a Cement Kiln",
in L. C. Blayden and S. G. Epstein, Spent Potlining Symposium,
JOM 36 (1984) 7, 22-32.
Lit. 16-32: V. R. Spironello, R. H. Nafziger,

"An Evaluation of Used Alunium Smelter Potlining as a Substitute for
Fluorspar",
US Bur. Mines Rep. RI 8530 1981.

"Treatment of Spent Potlining – an Update, Part I",
Aluminium 1997, 73, 51-52.

"Treatment of Spent Potlining - an Update, Part II",
Aluminium 1997, 73, 159-162.

"Treatment of Spent Potlining - an Update, Part III",
Aluminium 1997, 73, 248-250.

16-22
17.1. Summary................................................................. 17-3
17.2. Introduction ............................................................ 17-3
17.3. Navigating through ElyseSem .............................. 17-4

17.3.1. Navigation Panel.......................................................... 17-5
17.3.2. Page Navigation Toolbar.............................................. 17-7
17.4. Searching Information ........................................... 17-7

17-1

17-2
17.1. Summary
The contents of the ALCAN Electrolysis Seminar Textbook are issued

in the Adobe Acrobat format to be readable on the screen of a per-
sonal computer (ElyseSem). The use of ElyseSem is described, e.g.
how to navigate through the document and how to retrieve the
wanted information.
17.2. Introduction
ElyseSem is the contents of the ALCAN Electrolysis Seminar Textbook

issued in the Adobe Acrobat format to be readable on the screen of a per-
sonal computer.
Figure 17-1: First Page of ElyseSem. The content of the ALESA Electrolysis Seminar is
issued in the Adobe Acrobat format and can be read on the computer screen
using an Adobe Reader.

17-3
To read ElyseSem you need an Adobe Reader. If a reader is not already

available on your machine you may install the program from the Seminar
CD (read please the info.txt file on the CD) or from the Adobe-Website
(http://www.download.com/Adobe-Reader/3000-2378_4-10000062.html).
You find on the CD the Adobe Reader 7.0 and the new version 8.0 for
Windows XP (Service Pack 2) and Vista.
Although the use of the Adobe Reader is a standard procedure the follow-
ing pages explain some basic features of the software.
17.3. Navigating through ElyseSem
You can add to the reader interface shown in Figure 17-1 a menu bar, tool
bars and a navigation panel.
Figure 17-2: The Menu Bar, the Tool Bars and the Navigation Panel of the Adobe
Reader.
You do this by activating the corresponding menu items or by pressing the

F8 to toggle the toolbars and F9 to show and hide the menu bar.

17-4
17.3.1. Navigation Panel

On the Navigation Panel you find among others the Pages Button, the
Bookmarks Button and the Help Button (Figure 17-3) .
Figure 17-3: Pages Button, Bookmarks Button and Help Button of the Navigation Panel.
If you press the Pages Button the Adobe Reader shows miniature
previews (thumbnails) of the pages (Figure 17-4). After Clicking on one of
the thumbnails you jump quickly to the corresponding representation of
this page in the pages window.
Activating the Bookmarks Button shows a Table of Contents that
corresponds to the table of the Electrolysis Seminar Textbook. Press the
plus (+) or minus (-) button to expand or collapse the bookmark content.
Clicking on one item of the Table of Contents the Adobe Reader shows
the corresponding page in the pages window.

17-5
Figure 17-3: After you press the Pages Button the reader shows pages thumnails.
Figure 17-4: After you press the Bookmarks Button the reader shows a Table of Contents
corresponding to the table of the Electrolysis Seminar Textbook.
And finally when you press the Help button you may call different help
types of the Adobe Reader.

17-6
Figure 17-4: After you press the Help Button you may call several types of help facilities
of the Adobe Reader.
17.3.2. Page Navigation Toolbar

The next figure (Figure 17-5) shows the buttons and the Pages Number
Input Field of the Page navigation toolbar. In this input filed you select a
page number (332) of the documents total number of pages (431) to be
shown in the pages window.
Figure 17-5: The buttons and the input field of the Page Navigation Toolbar.
With the Next/Previous Pages Button you scroll to the next (333) and pre-
vious (331) page of the actually shown page (333). With the Next/Previous
Page View Button you retrace your recent viewing path in the document.
17.4. Searching Information
You open the Find Toolbar using the menu: View→Toolbars →Find. Type
a search item (liquidus, for instance) in then Find Text Box and press the
Enter Button. The Adobe Reader find the next instance of the text. Press
the Find Next/Find Privious button to scroll to the next/previous place in
the document that contains the text.

17-7
Figure 17-5: Text search with the Find Toolbar.
When you press the Crtl+Shift+F keys or you select Edit→Search from the
Menu Bar the Search Window opens. After entering “liquidus” in the
search box and pressing the Search Button you receive the search results
corresponding to Figure 17-6. When you click on one of the lines in the
Results Box the corresponding page is shown in the Page Window.

17-8
Figure 17-5: Search results after searching for “liquidus” in ElyseSem.pdf.

17-9

17-10
18.1. Summary................................................................18-3
18.2. Components of the ElysePrg Main Window .......18-3

18.2.1. Introduction .................................................................18-3
18.2.2. Menu Bar ....................................................................18-4
18.2.3. Toolbar........................................................................18-6
18.2.4. Contents Pane ............................................................18-7
18.2.5. Calculation Pane.........................................................18-8
18.2.6. Status Bar .................................................................18-10
18.3. Plant Data Page ...................................................18-11

18.3.1. Overview ...................................................................18-11
18.3.2. Equations ..................................................................18-14
18.4. CO2/CO Gas, F-Compounds Production Page..18-17

18.4.1. Overview ...................................................................18-17
18.4.2. Equations ..................................................................18-19
18.5. Point Feeder Setting Page..................................18-25

18.5.1. Overview ...................................................................18-25
18.5.2. Data Plot ...................................................................18-30
18.5.3. Equations ..................................................................18-33
18.6. Aluminum Fluoride Feeding Page .....................18-37

18.6.1. Overview ...................................................................18-37
18.7. Bath Composition Page......................................18-40

18.7.1. Overview ...................................................................18-40
18.7.2. Plot Bath Composition, Properties Window...............18-41
18.8. Bath/Liquidus Temperature Page ......................18-44

18.8.1. Overview ...................................................................18-44

18-1
18.9. Physical Bath Properties Page.......................... 18-46

18.9.1. Overview .................................................................. 18-46
18.10. Pot Voltage Page .......................................... 18-48

18.10.1. Overview .................................................................. 18-48
18.10.2. Reference Current Intensity...................................... 18-51
18.11. Bath Voltage Page........................................ 18-56

18.11.1. Overview .................................................................. 18-56
18.11.2. Popup Windows ....................................................... 18-58
18.11.3. Equations ................................................................. 18-61
18.12. Energy Balance Page................................... 18-62

18.12.1. Overview .................................................................. 18-62
18.12.2. Equations ................................................................. 18-64
18.13. Anode Table Page ........................................ 18-68

18.13.1. Overview .................................................................. 18-68
18.13.2. Detemination of the Anodic Fanning Factor ............. 18-70
18.14. List Examples and Exercises Window ....... 18-73

18.14.1. Overview .................................................................. 18-73
18.15. Overview Plot Window................................. 18-75

18.15.1. Overview .................................................................. 18-75
18.16. Literature....................................................... 18-76

18-2
18.1. Summary
The purpose of ElysePrg is to execute calculation concerning the

aluminum production by electrolysis. ElysePrg is thought to be used
together with ElyseSem e.g. when studying the examples or solving
the exercises of ElyseSem. Consequently you find in ElyseSem the
theoretical and practical basis (derivations and applications) of the
relations which ElysePrg is employing.
18.2. Components of the ElysePrg Main Window
ElysePrg executes the calculations on different pages of the calculation

pane. On the contents pane you select a calculation page to execute the
wanted calculations.
Figure 18-2-1: shows the components of the ElysePrg Main Window.
You use the components of the ElysePrg Main Window in the following
way:

18-3
Menu Bar: you use the items of the menu bar to save or print the results of
calculations, to change the layout of the ElysePrg Main Window, to navi-
gate through the calculation pages or to get help.
Toolbar: you use the buttons of the toolbar to quit the application, to
open/close the content pane and to navigate through the calculation
pages.
Contents Pane: you select the page of the calculation pane i.e. the type of
calculation by clicking on the corresponding item of the content pane.
Calculation Pane: you execute the actual calculations on different pages of

the calculation pane. You make the input of the start values and ElysePrg
shows then the calculation results.
Status Bar: the status bars display messages and useful information with-
out interrupting your work. The status bar indicates also the number of the
calculation page you are working on.
18.2.2. Menu Bar
You display pulldown menus when you click with the mouse pointer on a
menu bar item. ElysePrg opens three pulldown menus namely the File,
View and Help pulldown menu.
File Pulldown Menu

When you click with the mouse pointer on the File menu item of the menu
bar you open the File pulldown menu (Figure 18-2-2).
Figure 18-2-2: File pulldown menu.

ElysePrg does the following when you click on a File pulldown menu item:
Open: displays the standard Open Dialog Box, in which you can locate
and open a file where you have saved the calculation results.
Save: use this command to save the calculation results to its current name
and directory. When you save for the first time, ElysePrg displays the

18-4
standard Save As Dialog Box so you can name your file. If you want to
change the name and directory of an existing file before you save it,
choose the Save As command.
Save As: use this command to save and name the file where you want to
save the calculation results. ElysePrg displays the Save As dialog box so
you can name your file. To save a file with its existing name and directory,
use the Save command.
Print: use this command to print the calculation results. This command
presents a Print dialog box, where you may specify the range of pages to
be printed, the number of copies, the destination printer, and other printer
setup options.
Print Preview: use this command to display the pages as they would ap-
pear when printed. When you choose this command, the main window will
be replaced with a print preview window in which one or two pages will be
displayed in their printed format. The print preview toolbar offers you op-
tions to view either one or two pages at a time; move back and forth
through the document; zoom in and out of pages; and initiate a print job.
Print Setup: use this command to select a printer and a printer connection.
This command presents a Print Setup dialog box, where you specify the
printer and its connection.
Recent Files: opens a popup menu containing the last four documents you
closed. Choose the file name you want to open.
Exit: use this command to end your ElysePrg session. You can also press
the Exit Button of the toolbar.
View Pulldown Menu

When you click with the on the View menu item of the menu bar you open
the View pulldown menu (Figure 18-2-3).
Figure 18-2-3: View pulldown menu.

ElysePrg does the following when you click on a View pulldown menu
item:

18-5
Toolbar: use this command to show or hide the toolbar. The toolbar menu
item is checked when the toolbar is visible.
Status Bar: use this command to show or hide the status bar. The status
bar menu item is checked when the status bar is visible.
Contents: use this command to open or close the contents pane. The con-
tents pane menu item is checked when the contents pane is opened. You
can also press the Contents Button of the toolbar to toggle the contents
pane on and off.
Next Page: use this command to show the next page of the calculation
pane. You can also press the Next Page Button of the Toolbar.
Previous Page: use this command to show the previous page of the calcu-
lation pane. You can also press the Previous Page Button of the tool-
bar.
Help Pulldown Menu

When you click with the on the Help menu item of the menu bar you open
the Help pulldown menu (Figure 18-2-4).
Figure 18-2-4: Help pulldown menu.

The Help menu provide you assistance with ElysePrg. ElysePrg does the
following when you click on a Help pulldown menu item:
Open ElyseSem: Use this command to display the opening screen of Ely-
seSem. ElyseSem is the content of the ALCAN Electrolysis Seminar Text-
book transformed in the Adobe Acrobat format. You can read the content
of ElyseSem on you computer screen (see Chapter 17. How to Use Ely-
seSem).
About ElysePrg: Use this command to display the copyright notice, version
number and other information about ElysePrg.
18.2.3. Toolbar
The toolbar (Figure 18-2-5) contains several buttons and an input field to
place your comments.

18-6
Figure 18-2-5: Toolbar.

ElysePrg does the following when you activate the following buttons with
the mouse pointer:
press this button to end your ElysePrg session. You can also activate the
Exit menu item of the File pulldown menu.
press this button to open or close the contents pane. You can also activate
the Contents menu item of the View pulldown menu to toggle the contents
pane on and off.
press this button to show the next page of the calculation pane. You can
also activate the Next Page menu item of the View pulldown menu.
press this button to show the previous page of the calculation pane. You
can also activate the Previous Page menu item of the View pulldown menu.
You show or hide the toolbar by activating the Toolbar item of the View
pulldown menu. If the toolbar is visible the toolbar menu item is checked.
18.2.4. Contents Pane
The Contents pane displays in an hierarchical way a list of header items

and calculation page items (Figure 18-2-6).
Figure 18-2-6: Contents pane with expanded and collapsed header items and calculation
page items.
The header items have a list of calculation page items associated with
them. By clicking an header item, you can expand and collapse the asso-
ciated list of calculation page items. You select the calculation page dis-
played in the calculation plane by clicking on a calculation page item.

18-7
18.2.5. Calculation Pane
You execute the actual calculation on the pages of the Calculation pane.
You change the start values and ElysePrg shows the corresponding calcu-
lation results (Figure 18-2-7).
Figure 18-2-7: A calculation page of the Calculation pane with activated (white) and deac-
tivated (gray) input fields, a combo box and the More Info Button.
You find following items on the calculation page:
Input Fields: you introduce a value into the input field by using the key-
board. After you press the carriage return key ElysePrg executes the cal-
culation with this new value.
In Figure 18-2-7 ElysePrg has calculated the aluminum production
(2295.1 t) for one pots, a time period of one day, a current intensity of 300
kA and a current efficiency of 95%. The input field of the aluminum
production is grayed since it receives actually output of ElysePrg e.g. the
result of the calculations.
Suppose you want to calculate the current efficiency for a given aluminum
production. You click on the greed aluminum production input field which
becomes white i.e. activated and you introduce a value for the aluminum
production, 2300 t for instance. ElysePrg determines then the current effi-
ciency (96.19 %) in the grayed i.e. deactivated current efficiency input
field.
Remark: You can change the state of an input field either by a double or
by a single mouse click. After clicking twice into the field it be-
comes grayed e.g. it receives calculation results. Clicking once
you bring the field back to its white input state.

18-8
Figure 18-2-8: Calculation of current efficiency for a given aluminum production.

Combo Boxes: if you clicking on a combo box with the mouse pointer a
popup window opens. You select predefined values, for instance the units
of the time period in Figure 18-2-9, from this popup list.
Figure 18-2-9: Popup list of the time period combo box.

Buttons: the calculation page may contain several types of buttons. When
you press the More Info Button you show or hide additional calculation re-
sults (Figure 18-2-10).
Figure 18-2-10: After pressing the More Info Button the calculation page shows more cal-
culated values. Pressing the button again hides these values.

18-9
18.2.6. Status Bar
The status bars displays messages and useful information without inter-
rupting your work. The message line on the status bar can display informa-
tion about program status, about a toolbar button or menu item. The status
bar displays details of what buttons and menu commands do as you move
the mouse cursor over them.
The status bar indicates also the number of the calculation page you are
working on.
In Figure 18-2-11 the mouse pointer is over the Contents Button. Corre-
spondingly the message line of the status bar shows informations about
the Contents Button.
Figure 18-2-11: Information about the Contents Button in the message line of the status
bar. The status bar shows also the page number of the calculation page.

18-10
18.3. Plant Data Page
18.3.1. Overview
On this page you find the input fields for the plant data: number of
pots, time period, current intensity, current efficiency and aluminum
production. You determine either one of these values using the other
data as input.
The determination of the aluminum production or current efficiency are

the standard cases. When clicking on their input fields ElysePrg toggles
them Grey e.g. receives calculation result or white e.g. it receives user
input.
If you want to calculate the other values you have to double click on the
input field with the mouse pointer.
Example 18-3-1: Determination of the aluminum production (Fig. 18-3-1).
Suppose you want to calculate the aluminum production for

Number of Pots: 120,
Time Period: 1 month,
Current Intensity: 308.462 kA,
Current Efficiency: 95.46 %.
Figure 18-3-1: Calculation result corresponding to Exp. 18-3-1.

Example 18-3-2: Determination of the current efficiency (Figure 18-3-2).
Suppose you want to calculate the current efficiency for

18-11

Aluminum Produc- 8400 t.
tion:
Please click on the Aluminum Production input field with the mouse pointer
to activate it. The Aluminum Production input field becomes white and the
Current Efficiency input field is now gray e.g. it receives the calculation re-
sult. Enter now 8400 and press the carriage return key.
Figure 18-3-2: Calculation results corresponding to Exp. 18-3-2.
Example 18-3-3: Determination of the current intensity (Figure 18-3-3).
Suppose you want to know how much you have to change the current in-
tensity to produce 8500 tons of aluminum with the following conditions:

Current Efficiency: 95 %.
Please click twice on the Current Intensity input field to deactivate it.
Change then the values in the corresponding input fields.

18-12
Figure 18-3-3: shows the calculation page corresponding to Exp. 18-3-3.

In a similar way you determine the time period and the number of pot by
first double click on the corresponding input field to deactivate it and the
making the necessary input.
Example 18-3-4: Determination of the time period (Figure 18-3-4).

Determine the time period with the following conditions:

Current Intensity: 310 kA,
Current Efficiency: 95 %.
Aluminum Production: 10000 t.

18-13
Example 18-3-5: Determination of the number of pots (Figure 18-3-5).

Suppose you want to calculate the number of pots to produce the 10000 t
of aluminum according to Exp. 18-3-4 in one month.

Current Efficiency: 95 %,
Aluminum Production: 10000 t.
18.3.2. Equations
ElysePrg calculates the Aluminum Production using equation 2-16 of

Chap. 2-5 modified with the number of pots:
PAl = 0.3356 ◊ nPots ◊ I ◊ t ◊ h (18-3-1)
You find more information about fractional current efficiency in Chap. 2-4.

18-14
The other values (number of pots, time period, current intensity and cur-
rent efficiency) are determined after rearranging Equ. 18-3-1.
If you press the More Info Button ElysePrg (Fig. 18-3-6) shows or hides
several additional values, which ElysePrg calculates according to Chap. 2.
Figure 18-3-6: Plant Data Page after you have pressed the More Info Button.
ElysePrg uses the following equations:
Chap. 2-3 Theoretical Production and Consumption: Equ. 2-5 to 2-8,

Chap. 2-4 Electrolytic Production and Consumption: Equ. 2-14 to 2-18,
Chap. 2-5 Specific Electrolytic Production and Consumption: Equ. 2-19 to 2-22.
If you press the Carbon/Inert Anodes Button you can toggle between the
mass balance of an electrolysis process that uses carbon anodes ( Fig.
18-3-6) or that uses non-consumable anodes ( Fig. 18-3-7) i.e. the an-
odes do not react with the anodic oxygen (see Chap 1.5).

18-15
Figure 18-3-7: Plant Data Page of Fig. 18-3-6 after you have pressed the Carbon/Inert
Anodes Button.
Dealing with inert anodes ElysePrg uses the following equations (see
Chap. 2.10):
Electrolytic Production of Oxygen: Equ. 2-45,
Specific Electrolytic Production of Oxygen: Equ. 2-47.

18-16
18.4. CO2/CO Gas, F-Compounds Production Page
18.4.1. Overview
On this page you determine the production of carbon dioxide and

carbon monoxide either in mass units (kg, t, etc.) or volume units
(m3). The gas volume depends on the gas temperature and on the
gas pressure.
This page shows also the masses of fluorine and fluorine com-
pounds (aluminum fluoride, cryolite) which are leaving the electro-
lytic pots with the pot gas.
Example 18-4-1: Volume Gas Production (Fig. 18-4-1).
Suppose you want to calculate corresponding to Exp. 18-3-1 how much

carbon dioxide and carbon monoxide at standard conditions is produced
for

Current Efficiency: 95.46 %,
Gas Pressure: 1.013 Bar,
Gas Temperature: 0° C.
For the fluorine compound production you must specify the bath composi-
tion and the bath temperature. The bath temperature is determined from
the liquidus temperature (Equ. 3-8 of Solheim) and the overheat:
Aluminum Fluoride: 12 %,
Calcium Fluoride: 4.5 %,
Aluminum Oxide: 2.5 %,
Lithium Fluoride: 0 %,
Magnesium Fluoride: 0 %,
Overheat: 10° C.

18-17
Figure 18-4-1: CO2/CO Gas, F-Compounds Production Page corresponding to

Exp. 18-4-1.
With the Carbon/Inert Anodes Button you can toggle between the car-
bon dioxide, carbon monoxide gas production of the conventional Hall-
Héroult process which uses carbon anodes ( Fig. 18-4-1) or that uses
non consumable anodes ( Fig. 18-4-2) i.e. the anodes do not react
with the anodic oxygen (see Chap 1.5).
Figure 18-4-2: CO2/CO Gas, F-Compounds Production Page of Fig. 18-4-1 after you
have pressed the Carbon/Inert Anodes Button.

18-18
If you press the More Info Button (Fig. 18-4-3) ElysePrg shows or hides
the parameters which it uses for its calculations.
Figure 18-4-3: CO2/CO Gas, F-Compounds Production Page of Fig. 18-4-1 after you
have pressed the More Info Button.
18.4.2. Equations
Mass Production: ElysePrg uses the equations 2-17 and 2-18 (Chap.
2.5) to calculate the mass production of carbon dioxide and carbon mon-
oxide (see Fig. 18-3-6):
Volume Production: to calculate the volume production of carbon dioxide

and carbon monoxide in dependence of temperature and pressure Ely-
sePrg combines the equation 2-28 (t0 = 0° C and p0 = 1.013 bar) with the
equations 2-30 and 2-31:

18-19
t + 27315
.
VPCO2 = 0.2091⋅ I ⋅ t ⋅ (2η − 1) ⋅ ⋅ 1013
.
(18-4-1)
27315
. p
t + 273.15 1013
VPCO = 0.4181◊ I ◊ t ◊ (1- h) ◊ ◊. (18-4-2)
273.15 p
Specific Volume Production: The equations for the specific consumption

correspond to the equations 2-32 and 2-33 with the modification according
to equation 2-28 (t0 = 0° C and p0 = 1.013 bar):
2η − 1 t + 27315
. .
SVPCO2 = 0.6230 ⋅ ⋅ ⋅ 1013 (18-4-3)
η .
27315 p
1− η t + 27315
.
SVPCO = 12460
. ⋅ ⋅ ⋅ 1013
.
(18-4-4)
η 27315
. p
Masses of Gaseous Species: The determination of fluorine compound

production assumes that the pot gas (carbon dioxide and carbon monox-
ide for conventional Hall-Héroult cells and oxygen for inert anode cells) is
saturated with the monomer NaAlF4, the dimer Na2Al2F8 and sodium fluo-
ride. In Chap. 3.12 you find the relations to determine the partial pres-
sures of the gaseous species in dependence of the bath composition and
the bath temperature.
Using the Ideal Gas equation 2-25 you write for the volume of the pot gas
and the moles of the monomer species:

18-20
R ⋅T
V = ( nCO2 + nCO ) ⋅
pGas
(18-4-5)
pM ⋅ V
nM =
R ⋅T
Putting these equation together and using Equ. 2-17 for the produced
mass of carbon dioxide and Equ. 2-18 for the produced mass of carbon
monoxide you write for the mass of the monomer NaAlF4:
nM ⎛⎜ EPCO2 EPCO ⎞
⎟ ⋅ M ⋅ pM
mM = = + M (18-4-6)
MM ⎜ MCO MCO ⎟ pGas
⎝ 2 ⎠
mass of the dimer Na2Al2F8:
⎛ EPCO EPCO ⎞ p
mD = ⎜ 2
+ ⎟ ⋅M ⋅ D
D (18-4-7)
⎜ MCO MCO ⎟ pGas
⎝ 2 ⎠
mass of sodium fluoride:

18-21
⎛ EPCO EP ⎞ pNaF
mNaF =⎜ 2
+ CO ⎟ ⋅M
NaF ⋅ (18-4-8)
⎝ 2 ⎠
and the total mass of the gaseous species:
mTotal = mM + mD + mNaF (18-4-9)
Specific Masses of Gaseous Species: You write the equation for the
specific masses of the gaseous fluorine species by replacing the electro-
lytically produced masses of carbon dioxide and carbon monoxide in the
Equ. 18-4-6 to 18-4-8 by their specific masses (Equ. 2-21 and Equ. 2-22).
Specific mass of the monomer NaAlF4:
⎛ SEPCO SEPCO ⎞
m(spec)M = ⎜ 2
+ ⎟ ⋅ M ⋅ pM (18-4-10)
⎜ MCO MCO ⎟ M pGas
⎝ 2 ⎠
Specific mass of the dimer Na2Al2F8:

18-22
⎛ SEPCO SEPCO ⎞ p
m(spec)D = ⎜ 2
+ ⎟ ⋅M ⋅ D
D (18-4-11)
⎝ 2 ⎠
Specific mass of sodium fluoride:
⎛ SEPCO SEPCO ⎞ pNaF

m(spec)NaF =⎜ 2
+ ⎟ ⋅M ⋅ (18-4-12)
⎜ MCO MCO ⎟ NaF pGas
⎝ 2 ⎠
Total specific mass of the gaseous species:
m(spec)Total = m(spec)M + m(spec)D + m(spec)NaF (18-4-13)
Fluorine Mass of Gaseous Species: You write for the fluorine mass of
sodium fluoride and the monomer NaAlF4 and dimer Na2Al2F8 gas species:
MF M M
mF = ⋅ mM + F ⋅ mD + F ⋅ mNaF
MM MD MNaF
mF = 0.6033 ⋅ ( mM + mD ) + 0.4525 ⋅ mNaF (18-4-14)
Mass of Cryolite and Aluminum Fluoride: The gaseous species react to

form cryolite and aluminum fluoride:

18-23
3 ( NaF ) ( AlF3 ) = (NaF )3 ( AlF3 ) + 2AlF3

3NaAlF4 = Na3 AlF6 + 2AlF3 (18-4-15)
Sodium fluoride reacts with aluminum fluoride to cryolite:
3NaF + AlF3 = (NaF )3 ( AlF3 )

3NaF + AlF3 = Na3 AlF6 (18-4-16)
Since the molecular weight of aluminum fluoride is twice the molecular

weight of sodium fluoride you write for the masses of aluminum fluoride
and cryolite:
4 2
mAlF = ⋅ ( mM + mD ) − ⋅ mNaF
3 9 3
(18-4-17)
5 5
mCry = ⋅ ( mM + mD ) + ⋅ mNaF
9 3

18-24
18.5. Point Feeder Setting Page
18.5.1. Overview
You determine on this page the point feeder setting using as start
values the percent alumina consumption, the alumina concentration
gradient or the start and end values of the alumina concentration.
You may also investigate the influence of the point feeder setting on
alumina concentration.
The following examples use the values:

Current Efficiency: 94.38 %,
Number of Point Feeders: 2
Mass of Alumina per Point Feeder 1.65 kg,
Shot:
Bath Mass : 3600 kg,
Specific Alumina Consumption: 1.95 kg/kg.
Remark: You set the Current Intensity and the Current Efficiency values
on the Plant Data Page (Page 1) of ElysePrg. The other values
of Page 1 like the Number of Pots or the Time Period are not
relevant for the following examples.
Figure 18-5-1:Values of the Plant Data Page (Page 1) for the point feeder examples.
You check these values on this Point Feeder Setting Page by pressing the
More Info Button.

18-25
Example 18-5-1: Point Feeder Setting from Percent Alumina Consumption

(Fig. 18-5-2).
Suppose you want to determine the point feeder setting (feeding frequency
e.g. the number of point feeder shots in one hour) to add alumina corre-
sponding to 120 % of alumina consumption.
Figure 18-5-2: Point Feeder Setting Page corresponding to Exp. 18-5-1.

18-26
You get some additional output on this page:
• click on to toggle between the

feeding frequency: or the
feeding time period:
• ElysePrg calculates the concentration gradient cAl2O3/dt e.g. the change of
alumina concentration in one hour,
• ElysePrg determines also the alumina concentration (6.25 % in Fig. 18-5-2)

which will be reached after a time period (180 min in Fig. 18-5-2) starting from
5.0 % and using the concentration gradient 0.51 %/h.
Example 18-5-2: Point Feeder Setting from Al2O3 Concentration Gradient

(Fig. 18-5-3).
In this example you want to determine the point feeder setting to add alu-
mina corresponding to a alumina concentration gradient of -0.5 %/h.

18-27
Remark: The time period to determine the end concentration was

changed to 120 min. Consequently ElysePrg calculates an end
alumina concentration of 4.0 %.
You may also want to calculate the time period to reach a given
(end)concentration, 2.0 % for instance. You click twice into the Time Pe-
riod input filed it becomes gray i.e. it receives the result of calculation.
You change now the value of the end concentration value to 0.2 %. After
pressing the carriage return key ElysePrg determines the time period:
Example 18-5-3: Point Feeder Setting from Start/End Al2O3 Concentration

(Fig. 18-5-4).
Suppose you want to increase the alumina concentration from 9.0 % to

14.5 % in five hours. What is the point feeder setting?
Similar to Example 18-5-2 you click once into the grayed End cAl2O3
and/or Time Period input fields to activate them for input. After the input of
the values (Start cAl2O3: 9.0 %, EndAl2O3: 14.5 % and Time Period:
300 min) ElysePrg calculates the point feeder setting.

18-28
Example 18-5-4: Influence of Point Feeder Setting. (Fig. 18-5-5).
Suppose you want to change the point feeder setting according to Exam-
ple 18-5-1 from the point feeder frequency 33.13 shots per hour to 60 (1/h)
to see the influence on the other values like the consumption percents or
the concentration gradient.
More Info
The More Info Button shows and hides those values, which were set on
the Plant Data Page (Page 1).
Figure 18-5-6: More Info Values of the Point Feeder Setting Page.
ElysePrg calculates the practical alumina consumption using the specific

alumina consumption. If you want to calculate the specific consumption
from a known practical consumption value click on the input field of the
practical consumption which becomes activated and enter the value,
11000 t for instance. ElysePrg determines then the specific alumina con-

18-29
sumption, which is shown in its grayed input field (1.982 kg/kg). Also the
other values are adapted correspondingly.
Figure 18-5-7: Point Feeder Setting Page corresponding to Example 18-5-1 with the
changed values.
18.5.2. Data Plot
Figure 18-5-8 shows the Point Feeder Setting Page corresponding to Ex-
ample 18-5-3 with the Show Plot Button.
Figure 18-5-8: Point Feeder Setting Page with the Show Plot Button.

18-30
If you press this button you open the Plot Point Feeder Settings Window
(Fig. 18-5-9) which shows the plot of the Percent Consumption (Cons. in
%) and the alumina concentration (cAl2O3 in %) on a time scale (Time in
min).
Figure 18-5-9: Plot Point Feeder Settings Window.
Dragging
If you change values in the Pointfeeder Setting Page of ElysePrg the cor-
respondingly changed values are plotted in the Plot Point Feeder Setting
Window. You may however also change values in the plot window by
dragging the corresponding plotted points with the mouse pointer. If, for
instance, you want to know the point feeder setting to change the alumina
concentration form 12 % to 6 % in say 200 minutes you drag first the point
on the ordinate of the cAl2O3 - Time plot to 12 % and the other point to
6 %. The Percent Consumption - Time diagram follows your dragging ac-
tion as the Pointfeeder Setting Page of ElysePrg. They show the new cal-
culated values corresponding to the new concentration settings
(Fig. 18-5-10).

18-31
Figure 18-5-10: Plot Point Feeder Settings Window after dragging.
Changing Resolution
You may also change the resolution of the plot axes by clicking on the ar-
rows of the axes with the mouse pointer and pressing the left mouse but-
ton to increase the resolution and the right mouse pointer to decrease the
resolution (Fig. 18-5-11).
Figure 18-5-11: Plot of Fig. 18-5-10 with changed resolution of the coordinate axes.

18-32
18.5.3. Equations
The Specific Alumina Consumption is defined as:
c Al O (18-5-1)
SC Al2O3 = 2 3
PAl
With equation 2-16 the Alumina Consumption in One Hour is:
C(1h) = 0.3356 ⋅ I ⋅ η ⋅ SC Al2O3 (18-5-2)
Using these equations different cases are treated:
Point Feeder Setting determined from Percent Alumina Consumption

(see Example 18-5-1):
The point feeder setting (e.g. the alumina feeding frequency) is given with
F(%) ⋅C(1h) (18-5-3)

f PF =
100 ⋅ n PF ⋅ S PF
and the alumina concentration gradient with:
dc =
cF (%) h
− 100 ⋅C (1h)
(18-5-4)
MB

18-33
Knowing the concentration gradient the end alumina concentration is de-

termined form the start alumina concentration and the time period:
c End = c Start + dc ⋅ t c (18-5-5)
or the time period is calculated for the start and end alumina concentra-
tions:
c End − c Start
tc = (18-5-6)
dc
Point Feeder Setting determined from Concentration Gradient:

The point feeder setting (e.g. the alumina feeding frequency) is given with
dc ⋅ M B
+ C(1h)
f PF = 100 (18-5-7)
n PF ⋅ S PF
and the Percent Alumina Consumption:
100 ⋅C (1h) + dc ⋅ M B
F(%) = (18-5-8)
C(1h)

18-34
Please use Equations 18-5-5 and 18-5-6 to determine the end alumina
concentration or the time period.
Point Feeder Setting determined from Start/End Alumina Concentra-

tion and Timeperiod
Equation 18-5-6 is changed to determine the concentration gradient.
c End − c Start
dc = (18-5-9)
tc
The concentration gradient of the Equations 18-5-7 and 18-5-8 are used to
calculate the other values.
Influence of Point Feeder Setting

If you know the point feeder setting you find the percent alumina consump-
tion by rearranging Equations 18-5-3:
100 ⋅ fPF ⋅ nPF ⋅ SPF

F(%) = (18-5-10)
C (1h)
The other values you find according to the case of point feeder setting
from consumption.

18-35
Practical Alumina Consumption

(see More Info):
With equations 2-16 the practical alumina consumption is determined from

the specific alumina consumption with:
C Al O = 0.3356 ⋅ nPots ⋅ I ⋅ t ⋅ η ⋅ SC Al O (18-5-11)

2 3 2 3
and vice versa the specific alumina consumption from the practical alu-
mina consumption with:
C Al O
SC Al O = 2 3
(18-5-12)
2 3 0.3356 ⋅ nPots ⋅ I ⋅ t ⋅ η

18-36
18.6. Aluminum Fluoride Feeding Page
18.6.1. Overview
Components of the Aluminum Fluoride Consumption

To keep the aluminum fluoride content of the electrolyte constant an alu-
minum fluoride feeder adds aluminum fluoride from time to time. This alu-
minum fluoride consumption is due to several sinks and sources:
• Sodium Oxide: aluminum fluoride reacts with the sodium oxide of alu-
mina to form cryolite and aluminum oxide (see Exer-
cise 2-6),
• Carbon Blocks: the carbon lining of the electrolytic pot absorbs sodium
fluoride. These carbon blocks are a source or sink of
aluminum fluoride. Absorption of sodium fluoride in-
creases the aluminum fluoride content in the electrolyte.
Correspondingly the desorption of sodium fluoride de-
creases the aluminum fluoride concentrations.
• Ledge, Sludge: there are other sources (or sinks) of aluminum fluoride:
the composition of solid ledge may be different to the
electrolyte similar as the composition of bottom sludge.
Melting of ledge or dissolution will change the content of
aluminum fluoride in the electrolyte.
• Charged Alumina: in the scrubber alumina adsorbs fluoride compounds
which will change the electrolyte composition when
alumina is dissolved.
ElysePrg takes care of these factors in the following way (see Fig. 18-6-1):
• you specify the sodium oxide content of alumina. ElysePrg calculates the
corresponding alumina fluoride consumption using the alumina consumption
of the Point Feeder Setting Page (see Chap. 18.5.1).
• you may either specify the consumption of the contributors like carbon
blocks, ledge, bottom sludge, alumina and ElysePrg calculates the total
aluminum fluoride consumption,
• or you enter the value of total aluminum fluoride consumption and ElysePrg
calculates the consumption value for the other consumers.
You make your selection by clicking on that field where you want to make
the input. You may also double click on an input field i.e. it receives then
the calculation result (see Chap. 18.2.5).

18-37
Aluminum Fluoride Feeding Page
Figure 18-6-1 shows the Aluminum Fluoride Feeding Page.
Figure 18-6-1: Aluminum Fluoride Feeding Page.
Figure 18-6-1 shows that the sodium oxide content in alumina is set to
0.5 %. The aluminum fluoride consumption according to this sodium con-
tent is 17.6 g/kg where ElysePrg has used an alumina consumption of
1.95 kg Al2O3 per kilogram aluminum. The user has set the total consump-
tion of aluminum fluoride to 25 g aluminum fluoride per kilogram aluminum
hence the consumption of the other contributors is 7.4 g AlF3 per kg alu-
minum.
Similar to the other ElysePrg pages you can show or hide the additional
information by pressing the More Info Button.
On the left side Figure 18-6-1 shows the the values and parameters which
Elyseprg is using to calculate the AlF3 consumption. If you know the con-
sumption of aluminum fluoride you calculate similar to the alumina con-
sumption (see Chap. 18.5.1) the specific aluminum fluoride consumption
by clicking on the Practical AlF3 Consumption input field and specifying the
practical consumption value.

18-38
On the right side ElysePrg determines the cryolite production correspond-

ing to the sodium oxide content in the alumina. If you know your actual
cryolite production (Practical Cryolite Production) ElysePrg calculates the
cryolite loss.
Options
As on the other calculation pages you have several options to calculate
the aluminum fluoride point feeder setting:
• Consumption Units: you may choose between gram aluminum fluoride per
kilogram aluminum or kilogram aluminum fluoride per
day by clicking with the point feeder on the on the units,
• Consumption: you may choose between the input of the total con-
sumption or the contributors like side ledge, sludge and
carbon bottom blocks,
• Feed: you may specify the aluminum fluoride feed and
ElysePrg calculates the change of the aluminum fluoride
concentration in a given time period or you specify the
concentration change and the feed is calculated.

18-39
18.7. Bath Composition Page
18.7.1. Overview
Figure 18-7-1 shows the Bath Composition Page.
Figure 18-7-1: Bath Composition Page.
Before you enter the concentration values you have to chose the Standard
System:
Sodium Cryolite - Aluminum Fluoride - Calcium Fluoride - …
or the system:
Sodium Fluoride - Aluminum Fluoride - Calcium Fluoride - …
and the units (see Tab 3-1):
Mass Fraction (Weight Percent)
or
Mole Fractions (Mole Percent).

18-40
You may enter the aluminum fluoride also as Bath Ratio when using
weight percent or as Cryolite Ratio when using mole percent.
If the alumina concentration is less than the alumina anode effect concen-
tration ElysePrg shows the anode effect sign and locks the input/output
operation of the Pot Voltage Page (Figure 18-7-2).
Figure 18-7-2: Bath Composition Page showing the anode effect sign.
EWR is the Equivalent Ratio (see also the Examples 3-1, 3-2, and 3-3).
18.7.2. Plot Bath Composition, Properties Window
After pressing the Show Plot Button of the Bath Composition Page a win-
dow pops up which shows a graphical representation of the concentrations
values of the bath components and of some bath properties like bath and
liquidus temperature, electric conductivity and the bath densities together
with the density of aluminum (Fig. 18-7-3).

18-41
Figure 18-7-3: Bath Composition, Properties Window corresponding to the values of

Fig. 18-7-1.
Similar to the Plot Point Feeder Settings Window you may change the bath
composition i.e. the concentration value of a component by dragging the
rectangle of the plot to the new value. This was done in Fig 18-7-4 for the
aluminum fluoride concentration which was changed from 12 % if Fig. 18-
7-3 to 4.5 % in Fig. 18-7-4. The other values have changed corresponding
to the new bath composition.
You may change the resolution of the concentration ordinate by clicking on
the coordinate arrow and holding down the left mouse button to increase
and the right mouse button to decrease the resolution.
Figure 18-7-4: Bath Composition, Properties Window after dragging the aluminum fluo-
ride concentration to 4.5 %.

18-42
You may change the property ordinate by clicking on that value area which
scale you want to have plotted. Figure 18-7-5 shows the Bath Composi-
tion, Properties Window after changing the concentration ordinate resolu-
tion by clicking with the mouse pointer on the ordinate arrow and holding
down the right mouse button.
After clicking on the density area of the plot the properties ordinate shows
the scale for the density values.
Figure 18-7-5 Bath Composition, Properties Window with density values scale.

18-43
18.8. Bath/Liquidus Temperature Page
18.8.1. Overview
You determine on this page the liquidus temperature of bath of given

composition after having selected the equation of the liquidus tem-
perature. Setting the overheat (or temperature) ElysePrg calculates
the bath temperature (or overheat).
If you know the liquidus temperature ElysePrg calculates the corre-
sponding aluminum fluoride concentration supposing that the con-
centration of the other bath components are known.
Liquidus Temperature
Corresponding to Example 3-4, for instance, ElysePrg calculates the liq-
uidus temperature after you have chosen the relation for the liquidus tem-
perature (see Chap 3.6) for a given bath composition (13.3 wt% AlF3, 4.7
wt% CaF2, 2.7 wt% Al2O3, 1.3 wt% LiF and 0.7 wt% MgF2 in Fig. 18-8-1).
Figure 18-8-1 shows the Bath/Liquidus Temperature Page corresponding
to Exp. 3-4.
Figure 18-8-1: Bath/Liquidus Temperature Page corresponding to Exp. 3-4.
Aluminum Fluoride
If you click with the mouse pointer on the liquidus input field it becomes
white. If you change the value of the liquidus temperature ElysePrg calcu-
lates the aluminum fluoride concentration using the equation of Solheim
(Equ. 3-8) keeping the other concentration values. In Fig. 18-8-2 ElysePrg
has calculated the aluminum fluoride content for a liquidus temperature of
950 °C.

18-44
Figure 18-8-2: Bath/Liquidus Temperature Page after the aluminum fluoride concentra-
tion was calculated from the liquidus temperature.
With the liquidus temperature you determine the bath temperature using a
constant overheat or you use a constant bath temperature and ElysePrg
calculates consequently the overheat.
In the same way ElysePrg determines the liquidus temperature according
to the bath composition and your choice of equation or ElysePrg calculates
the aluminum fluoride concentration using the liquidus temperature, the
concentrations of the other bath components and the equation of Solheim
(Equ. 3-8).
Show Plot
The Show Plot Button shows and hides the Plot Bath Composition, Prop-
erties Window (Chap. 18.7.2).

18-45
18.9. Physical Bath Properties Page
18.9.1. Overview
On this page you determine several physical properties of the elec-

trolytic bath for a given bath composition and bath temperature. Your
input on this page is to select the equation for some properties.
The input for the bath composition you have made on the Bath Composi-
tion Page or on the Bath Composition, Properties Window. The bath tem-
perature you have determined on the Bath/Liquidus Temperature Page.
On this page you choose the relation ElysePrg uses to calculate the elec-
tric conductivity and the densities of aluminum and the electrolyte. Fig-
ure 18-9-1 shows the Physical Bath Properties Page corresponding to
Exp. 3-5.
Figure 18-9-1: Physical Bath Properties Page corresponding to Exp. 3-5.
Remark: By clicking on Electric Conductivity you toggle between the out-

put of electric conductivity and resistivity.
ElyseSem treats the bath properties in several chapters and examples:
Electric Resistivity: Chap. 3.7, Exp. 3-5,

Densities: Chap. 3.8, Exp. 3-6,
Alumina Solubility: Chap. 3.9, Exp. 3-7,
Viscosities: Chap. 3.10, Exp. 3-8,
Vapor Pressure: Chap. 3.12.

18-46
More Info
The More Info Button shows and hides those values, which you have se-
ton the Bath Composition Page and ElysePrg has calculated on the
Bath/Liquidus Temperature Page (Fig. 28-9-2). The page shows also the
total vapor pressure and the partial pressures of the gaseous species (see
Chap. 3.12).
Figure 18-9-2: More Info Values Physical Bath Properties Page.
Show Plot
The Show Plot Button shows and hides the Plot Bath Composition,
Properties Window (Chap. 18.7.2).

18-47
18.10. Pot Voltage Page
18.10.1. Overview
On this page the pot voltage is split up into several components (see
Chap. 4.2). You find the input fields for the bath voltage, pot voltage
and the interpolar distance. You determine either one of these values
using the other data as input.
Example 18-10-1: Input Interpolar Distance (Fig. 18-10-1).

If the input field of the interpolar distance is gray you click once with the
mouse pointer on it. The field becomes white and you enter the value of
the interpolar distance. ElysePrg calculates then the pot and bath voltage
using the new interpolar distance value.
The folowing calculation use corresponding to Exp. 18-10-1 the following

parameters:
Entry to Anode Voltage: 0.250 V,

Bottom (Cathode) Voltage: 0.350 V,
Collector Bars to Exit Voltage: 0.250 V,
Anode to Cathode Distance: 4.5 cm,
Bath composition: AlF3:13.3 %,
CaF2: 4.7 %,
Al2O3: 2.5 %,
LiF: 1.3 %,
MgF2:0.7 %,
Overheat: 10 °C constant,
32 Anodes: length: 1450 mm, width: 540 mm.
Bubble Layer Thickness: 0.5 cm,
Surface Coverage: 0.5,
Anodic Fanning Factor: 1.1,
Cathodic Fanning Factor: 1.2,
Liquidus Temperature: relation of Solheim (Equ. 3-8),
Electric Conductivity: relation of Híveš (Equ. 3-9).

18-48
Figure 18-10-1: Pot Voltage and Bath Voltage calculated from the Interpolar Distance
(Exp. 18-10-1) .
Example 18-10-2: Input Bath Voltage (Fig. 18-10-2).
If you want to change the bath voltage click on the input field with the
mouse pointer. The field becomes white and you may change the values
of the bath voltage, for instance to 3.5 V. ElysePrg calculates then the pot
voltage and the interpolar distance using the new bath voltage value.
Figure 18-10-2: Pot Voltage and Interpolar Distance calculated from the Bath Voltage
(Exp. 18-10-2) .

18-49
Example 18-10-3: Input Pot Voltage (Fig. 18-10-3).

If you want to know the bath voltage and interpolar distance for a given pot
voltage, 4.2 V for instance you proceed in a similar way. Click on the input
field of the pot voltage with the mouse pointer and change the value.
ElysePrg calculates the bath voltage and the interpolar distance using the
new pot voltage value.
Figure 18-10-3: Bath Voltage and Interpolar Distance calculated from the Pot Voltage
(Exp. 18-10-3).
Example 18-10-4: Input Current Density (Fig. 18-10-4).

Suppose you want to execute the calculations for a known current density,
for instance 1.0 A/cm2. You click on the input field of the current density
with the mouse pointer and change the value. If you want go back and
have the current density being calculated you click twice on the input field.
It becomes gray and the old calculated value is shown.
Next you have to choose you calculation path according to the examples
18-10-1 to 18-10-3. Suppose you want to do the calculation for an interpo-
lar distance of 5.0 cm. You proceed like in Exp. 18-10-1.

18-50
Figure 18-10-4: Input current density (Exp. 18-10-4) .

Remark: It may happen that you get in the current density line a warning
like . It means that on the Anode Table Page ElysePrg has
determined a current density which was out of the limits. Ely-
sePrg has set the value of the current density to 0.7 A/cm2.
Click on the warning with the mouse pointer to go to the Anode
Table Page.
More Info
In Fig. 18-10-4 the user has pressed the More Info Button. ElysePrg
shows the ohmic and electrochemical components of the bath voltage.
These values are calculated on the Bath Voltage Page.
18.10.2. Reference Current Intensity
You may want to investigate the influence of changing the current intensity
on the other voltages. You declare a current intensity as the reference
value by clicking twice into the Reference input field. The Reference input
field and the voltages fields become grey i.e. the values are calculated by
ElysePrg. If you change now the current intensity the voltage values are
calculated correspondingly keeping the resistances of the current conduc-
tors constant.
Example 18-10-5: Change current intensity (Fig. 18-10-5).
Suppose you want to investigate what is happening if you increase in Ex-
ample 18-10-1 the current from 185 kA to 200 kA keeping the interpolar
distance constance.

18-51
After clicking on the Reference input field twice with the mouse pointer and
changing the current intensity to 200 kA ElysePrg calculates the answer.
Figure 18-10-5: Increase of current intensity from 185 kA to 200 kA using the Reference
Current Intensity.
If you click again into the Reference input field all the fields become white
and the values are reset corresponding to Fig. 18-10-1.
Popup Windows
Experience has shown that there are discrepancies between the users
how the pot voltage should be split up (see Chap. 4.2) . If you click on the
names of the components of the pot voltage ElysePrg shows in popup
windows how the program interprets these components. In Fig. 18-10-6
the user has clicked on the names of pot voltage component and has
opened the corresponding popup windows. You close a window by clicking
on the name again.

18-52
Figure 18-10-6: Pot Voltage Page after the user has clicked on the names of the compo-
nents and has opened the corresponding popup windows.
Figure 18-10-7 shows the Pot Voltage Popup Window. It displays how the
pot voltage is determined and contains the actual value of the pot voltage.
If the value of the pot voltage changes ElysePrg updates the value in the
Pot Voltage Popup Window.
Figure 18-10-7: Pot Voltage Popup Window (see Fig. 4-1).
In the following figures you find the other popup windows:

18-53
Figure 18-10-8: Entry to Anode Voltage Popup Window (see Fig. 4-3).
Figure 18-10-9: Bath Voltage Popup Window (see Fig. 4-4).
Figure 18-10-10: Bottom (Cathode) Voltage Popup Window (see Fig. 4-5).
Figure 18-10-11: Collector Bars to Exit Voltage Popup Window (see Fig. 4-6).

18-54
Figure 18-10-12: Interpolar Distance Popup Window.

18-55
18.11. Bath Voltage Page
18.11.1. Overview
On this page you determine how the components of the bath voltage
are calculated.
Figure 18-11-1 shows the Bath Voltage Page corresponding to (Exp. 18-
10-1). The following part of the chapter describes how you use the input
fields of this page to determine the bath voltage.
Figure 18-11-1: Example of the Bath Voltage Page corresponding to Exp. 18-10-1.
Current Intensity: in this field you change the current intensity value for
the Pot Voltage Page (Chap. 18.10) and the Anode Table Page
(Chap. 18.13).
Anodic Current Density: this input field shows the anodic current density,
which ElysePrg uses for the calculations. You may use the value ElysePrg
has calculated and that ElysePrg shows in the gray input field.
If you want however similar to Exp. 18-10-4 to execute the calculations for
a known current density you click on the input field of the current density
with the mouse pointer and change the value. If you want go back and
have the current density being calculated you click twice on the input field.
It becomes gray and the old calculated value is shown.

18-56
Remark: As mentioned on the Pot Voltage Page (Chap. 18.10) it may

happen that in the current density line you may find a warning
like . It means that on the Anode Table Page (Chap. 18.13)
ElysePrg has determined a current density which was out of the
limits. ElysePrg has set the value of the current density to 0.7
A/cm2. Click on the warning with the mouse pointer to go to the
Anode Table Page (Chap. 18.13).
Fanning Factors: in the input field of the anodic fanning factor you have
the choice to use the value calculated on the Anode Table Page (marked
with Haupin) or to use a value of your own (marked Input).
In the input field of the cathodic fanning factor you enter your value.
Remark: If you have selected to use a known current density i.e. you
have entered a value in the current density input field ElysePrg
deactivates the input fields for the fanning factors.
Electrochemical Bath Voltage (see Electrochemical Bath Voltage Win-

dow, Chap. 18-12.1): this input field shows the result of the calculation of
the electrochemical bath voltage.
If you want to use a known value you click on the input field of the Electro-
chemical Bath Voltage with the mouse pointer. The input field becomes
white and you may change the value. If you want go back and you want
that ElysePrg calculates the Electrochemical Bath Voltage you click twice
on the input field. It becomes gray and the old calculated value is shown.
Bubble Voltage: you may want that ElysePrg calculates the bubble volt-
age from the bubble layer thickness and the surface coverage. In this case
either you enter the values of the layer thickness and the surface coverage
into the input fields or you select the values calculated by Haupins equa-
tions (Equ. 18-11-1 and Equ. 18-11-2).
If you want to use a known value of the bubble voltage you click on the in-
put field with the mouse pointer. The input field becomes white and you
enter the value. ElysePrg has deactivated the input fields of the bubble
layer thickness and of the surface coverage. If you want go back and you
want that ElysePrg calculates the bubble voltage you click twice on the in-
put field. It becomes gray and the old calculated value is shown.
Ohmic Bath Voltage (see Ohmic Bath Voltage Window, Chap. 18.11.2):
ElysePrg shows the calculated ohmic bath voltage.
Similar to other pages the Bath Voltage Page contains a Carbon/Inert

Anodes Button. You toggle between the bath voltage values of the con-
ventional Hall-Héroult process which uses carbon anodes or of a pro-
cess that uses non consumable anodes (see Chap 1.5).

18-57
18.11.2. Popup Windows
Similar to the Pot Voltage Page ( Chap. 18-10) you may open popup win-
dows, which show details how ElysePrg has calculated the components of
the bath voltage. In Fig. 18-11-2 the user has clicked on some names of
the pot voltage components and has opened the corresponding popup
windows. You close a popup window by clicking on the name again.
Figure 18-11-2: Bath Voltage Page corresponding to Fig. 18-11-1 after the user has
clicked on the names of the components and has opened the corre-
sponding popup windows.
Current Density Window (see Chap. 4.5):

ElysePrg opens the Current Density Window after you have clicked on the
Anode Current Density name (Fig. 18-11-3).
On the left side you see the values of the anode and cathode surface
without fanning i.e. the values of the geometrical surface (Equ. 4-3) and on
the left side the values with fanning. This surface values depend on the
fanning factors you have chosen on the Bath Voltage Page.
ElysePrg calculates the Anodic and Cathodic Current Density using
Equ. 4-2 the Critical Current Density with Equ. 4-11 and the Limiting Cur-
rent Density with Equ. 4-10.

18-58
Figure 18-11-3: Current Density Window.
Electrochemical Bath Voltage Window

ElysePrg opens the Electrochemical Bath Voltage Window after you have
clicked on the Electrochemical Bath Voltage name (Fig. 18-11-4). On this
page you see the voltage components which sum up to the Electrochemi-
cal Bath Voltage.
Remark: If you have set the value of the Electrochemical Bath Voltage on
the Bath Voltage Page ElysePrg deactivates the Electrochemi-
cal Bath Voltage Window i.e. you CANNOT open it.
ElysePrg uses Equ. 4-1 to calculate the Reversible Decomposition Voltage

(see Chap. 4.5). It determines the Anodic Concentration Overvoltage with
Equ. 4-9 and the Anodic Reaction Overvoltage with Equ. 4-8. Finally
ElysePrg uses Equ. 4-12 to calculate the Cathodic Concentration Over-
voltage. For the calculation of the overvoltages see Chap. 4.7.
Figure 18-11-4: Electrochemical Bath Voltage Window.
If you have pressed the Carbon/Inert Anodes Button to switch to a process

that uses non consumable anodes ElysePrg changes the values of the
Electrochemical Bath Voltage Window according to Fig. 18-11-5.

18-59
Figure 18-11-5: Electrochemical Bath Voltage Window for an electrolysis process that
uses inert anodes.
On this window you find no Anodic Reaction Overvoltage value since the
oxygen, which the electrolysis process is producing does not react with the
anode. Also ElysePrg has changed the value of the Reversible Decompo-
sition Voltage according to the modified electrolysis equation (Equ. 5-9)
and the different Gibbs Free Energy (Equ. 5-11).
Ohmic Bath Voltage Window

ElysePrg opens the Ohmic Bath Voltage Window after you have clicked on
the Ohmic Bath Voltage name (Fig. 18-11-6). On this page you see the
parameters and components which ElysePrg uses to determine the Ohmic
Bath Voltage.
Figure 18-11-6: Ohmic Bath Voltage Window.

ElysePrg applies Equ. 4-7 to calculate the Bubble Voltage (see Chap. 4.6)
using the values of the Bubble Layer Thickness and Anode Surface Cov-
erage from the Bath Voltage Page.
ElysePrg detemines the Voltage Drop with Equ. 4-6, the Electric Conduc-
tivity of Electrolyte from the Physical Bath Properties Page (Chap. 18.9)
and the Interpolar Distance from the Pot Voltage Page (Chap. 18.10).
The Ohmic Bath Voltage is then the sum of the Bubble Voltage and the
Voltage Drop.

18-60
18.11.3. Equations
A recent paper (Lit. 18-2) cites relations to estimate the bubble layer thick-
ness and the gas surface coverage to calculate the bubble voltage (Equ.
4-7).
Bubble Layer Thickness (Lit. 18-2, Equ. 30):
0.5517 + j A
dBub = (18-11-1)
1+ 2.167 ⋅ j A
Surface Coverage (Lit. 18-2, Equ. 31):
ϕ = 0.5090 + 0.1823 ⋅ j A − 0.1723 ⋅ j A2 + 0.05504 ⋅ j A3 +

0.5517 + 0.3781⋅ R
+
1+ 1.637 ⋅ R
( )
AE (18-11-2)
0.431+ 0.1437 ⋅ cAl O − cAl
2 3 2 O3
( AE
1+ 7.353 ⋅ cAl O − cAl
2 3 2 O3
)
You find the values of the bath ratio (Equ. 3-3), the alumina content of the
electrolyte and the anode effect alumina concentration on the Bath Com-
position Page (Chap. 18.7).

18-61
18.12. Energy Balance Page
18.12.1. Overview
You determine on this page the enthalpies of the chemical reactions

to produce aluminum. Summing up these energy consumption with
the electrical energy input this page shows the heat which is lost to
the environment.
The following calculations use the values of example 4.2 (Alusuisse 180
kA pot):

Entry to Anode Voltage: 0.250 V,
Bottom (Cathode) Voltage: 0.350 V,
Collector Bars to Exit Voltage: 0.250 V,
Interpolar Distance: 5 cm,
Anodic Fanning Factor: 1.05,
Cathodic Fanning Factor: 1.0,
Bubble Voltage: 0.0 V,
Bath composition: AlF3: 10.2% (this value is changed from 11% to
10.2% to yield a bath temperature of 970 °C),
CaF2: 5.0 %,
Al2O3: 2.5 %,
LiF: 0.5 %,
MgF2:0.5 %,
Liquidus temperature: 960 °C relation of Solheim (Equ. 3-8),
Overheat: 10 °C constant,
Electrical conductivity: 2.2272 S/cm relation of Híveš (Equ. 3-9),
Example 18-12-1: Energy Balance of the Alusuisse 180 kA Pot for a cur-
rent efficiency of 95 %. Calculate the results in specific values i.e.
kWh/kgAl (Fig. 18-12-1).

18-62
Figure 18-12-1: Energy Balance Page corresponding to Exp. 18-12-1.

You change the units by selecting the them in the units field of the Total
Energy. By pressing the Show/Hide Info Button you toggle the output val-
ues from other pages of ElyseSem (Figure 18-12-2).
Figure 18-12-2: Energy Balance Page corresponding to Fig. 18-12-1 however the units in
kilo watts and the Show Info Button pressed.

18-63
Similar to other pages the Enrgy Balance Page contains a Carbon/Inert

Anodes Button. You toggle between the bath values of a conventional
Hall-Héroult electrolytic pot which uses carbon anodes or a pot that
uses non consumable anodes (Fig. 18-12-3, see Chap 1.5).
Figure 18-12-3: Energy Balance (corresponding to Fig. 18-12-1 or Fig. 18-12-2) of an

electrolytic pot that uses inert (non consumable) anodes.
The power value to heat the carbon anodes and to react carbon with the
anodic oxigen are lacking. Whereas the use of inert increases the bath
voltage i.e. the corresponding power values due to the changed reversible
decomposition voltage.
18.12.2. Equations
The calculation follows the discussion given in Chap. 5.3, Energy Re-
quirements. ElysePrg introduces the temperature dependence of the en-
thalpies by interpolating linearly between the temperatures 827 °C and
1027 °C (1100-1300 K) the values which were taken from the JANAF-
Tables (Lit. 18-1).

18-64
Enthalpies to heat alumina from 25 °C to t °C:
DH Al2O3 = 90.535 + 0.12823 ¥ t - 827 d i (18-12-1)
Enthalpies to heat the carbon anodes from 25 °C to t °C:
DH C = 13.989 + 0.002275 ¥ t - 827 e j (18-12-2)
Enthalpy of formation of alumina at t °C:
DH f ( Al2O3 ) = -1692.437 + 0.011235 ¥ t - 827 d i (18-12-3)
Enthalpy of formation of carbon dioxide at t °C:
DH f ( CO2 ) = -394.838 + 0.002095 ¥ t - 827 d i (18-12-4)
Enthalpy of formation of carbon monoxide at t °C:
DH f ( CO) = -112.586 + 0.00642 ¥ t - 827 d i (18-12-5)
The enthalpy of reaction according to Equ. 5-3:
1
Al2 O 3 +
3 3 1
C = Al + 2 − CO 2 +
FG
3 1
− 1 CO
IJ FG IJ
2 4η 4 η H
2 η K H K (18-12-6)
is then:

18-65
1 3 1 F I 3 1 F I
DHreact =
2 4 h GH JK
DH f ( Al2O3 ) - 2 - DH f ( CO2 ) -
2 h GH
- 1 DH f ( CO ) JK (18-12-7)
and the total energy necessary to produce aluminum:
1 3
DH tot = DHreact + DH Al2O3 - DH C (18-12-8)
2 4h
ElysePrg calculated the results of Exp. 18-12-1 (Fig.18-12-1) in the follow-

ing way:
Total Specific Energy (Equ. 5-2):

4.280
E= = 13.424 kWh/kg
0.3356 × 0.95
Specific Energy of Busbars (corresponding to Equ. 5-2):

0.5
= 1.568 kWh/kg
0.3356 × 0.95
Specific Energy into Pot (corresponding to Equ. 5-2):

3.784
= 11.866 kWh/kg
0.3356 × 0.95
Specific Energy to Heat Alumina (Equ. 18-12-6, Chap. 5.3):

1 1 108872
ΔH Al2O3 = × = 0.560 kWh/kg
2 2 26.9815 × 3600
Specific Energy to Heat Anodes (Equ. 18-12-6, Chap. 5.3):

3 3 17242
ΔHC = × = 0.140 kWh/kg
4η 4 × 0.95 26.9815 × 3600
Al2O3 Î Al + O2 (Equ. 18-12-6, Chap. 5.3):

1 1 1690823
ΔH f(Al2O3) = × = 8.704 kWh/kg
2 2 26.9815 × 3600

18-66
C+O2 Î CO2 + CO (Equ. 18-12-6, Chap. 5.3):

3⎛ 1⎞ 3⎛1 ⎞
⎜ 2 − ⎟ ΔH f(CO2) + ⎜ − 1⎟ ΔH f(CO) =
4⎝ η⎠ 2 ⎝η ⎠
3⎛ 1 ⎞
+ ⎛⎜ ⎞
395138 3 1 113504
⎜2 − ⎟× − 1⎟ × = 2.983 kWh/kg
4⎝ 0.95 ⎠ 26.9815 × 3600 2 ⎝ 0.95 ⎠ 26.9815 × 3600
Specific Energy of Reaction (Equ. 5-4):

ΔHreact = 8.704 − 2.983 = 5.721 kWh/kg
or
ΔHreact = ⎛⎜ ⎞
126098 1
+ 422961⎟ × = 5.721 kWh/kg
⎝ 0.95 ⎠ 26.9815 × 3600
Total Specific Energy Consumption by summing up

ΔH tot = 5.721 + 0.560 + 0.140 = 6.422 kWh/kg
or with Equ. 5-5

ΔH tot = ⎛⎜ ⎞
139030 1
+ 477397 ⎟ × = 6.422 kWh/kg
⎝ 0.95 ⎠ 26.9815 × 3600
or with Equ. 5-6:

ΔH tot = ⎛⎜ ⎞
1.431
+ 4.915 ⎟ = 6.422 kWh/kg
⎝ 0.95 ⎠
Specific Heat Loss by summing up:
Q = 11.868 − 6.422 = 5.434 kWh/kg

18-67
18.13. Anode Table Page
18.13.1. Overview
On this page you see the values to determine the Anodic Current
Density and to estimate the Anodic Fanning Factor with Haupin's
method.
Figure 18-13-1 shows the Anode Table Page corresponding to

Exp. 18-10-1.
Figure 18-13-1: Anode Table Page.
Current Intensity
In this input field you change the value of the current intensity for the Pot
Voltage Page (Chap. 18.10) and Bath Voltage Page (Chap. 18.11).
Anodic Current Density

According to Equ. 4-3 ElysePrg uses the Number, Length and Width val-
ues of the anodes to calculate the Geometric Anode Surface Area i.e.
the surface area of new anodes. With this value and the current intensity
ElysePrg determines the Geometric Anodic Current Density according
to Equ. 4-2 using a Fanning Factor of 1. The Geometric Current Density is
the current density of new anodes without any fanning.

18-68
In reality the electrolysis process consumes the anodes i.e. the surface
area is smaller than the geometric value. On the other hand the electric
current uses also the vertical sides of the anodes. To take account of
these effects ElysePrg uses the anodic Fanning Factor (Actually Used
Fanning Factor). On the Bath Voltage Page (Chap. 18.11) you choose to
use your own value (marked Input) or the value calculated by Haupin's
equations (marked Haupin). With this value ElysePrg calculates with
Equ. 4-2 the Electrolytic Anodic Current Density i.e. the anodic current
density which is actually used for the voltage calculations.
Set Anodic Current Density

On the Pot Voltage Page (Chap. 18.10) and the Bath Voltage Page (Chap.
18.11) you have the choice to execute the calculation with a known Elec-
trolytic Anodic Current Density. ElysePrg shows the Anode Table Page
according to Fig. 18-13-2.
Figure 18-13-2: Anode Table Page with set anodic current density (0.7 A/cm2).
As you see in Fig. 18-13-2 ElysePrg determines from the set current den-
sity a hypothetical anode surface (Anode Surface from Set Current Den-
sity) and deactivates the calculation of the Fanning Factor.
Current Density Out of Limits

It may happen that the calculated Geometric Anode Surface Area is out of
the limits i.e. a minimum value of 0.3 A/cm2 and a maxmum value of
2.5 A/cm2. In this case ElysePrg sets the Electrolytic Anodic Current Den-
sity to 0.7 A/cm2 (Fig. 18-13-3) and calculates the corresponding hypo-
thetical anode surface (Anode Surface from Set Current Density). In addi-
tion ElysePrg shows the warning on the Pot Voltage Page (Chap.
18.10) and on the Bath Voltage Page (Chap. 18.11).

18-69
Figure 18-13-3: Anode Table Page with anodic current density out of limits.
18.13.2. Detemination of the Anodic Fanning Factor
This chapter refers to the relations of Lit. 18-2 (Equ. 33 to 35). The algo-
rithm modifies the length and the width of every anode according to the
fanning which depends on the orientation of the anode face in the electro-
lytic pot. With this modified anode dimensions ElysePrg determines the
Effective Anode Surface.
Effective Anode Surface:

Equ. 18-13-1 shows the relation for the Effective Anode Surface:
Aeff = 10−6 ⋅ ∑ ( L + F1 + F2 ) ⋅ (W + F3 + F4 ) (18-13-1)

N
ElysePrg determines the average length and width of the anodes using the
surface area of the anodes butts (see Chap. 14.7.4) and the relation:

18-70
L = 100 + α ⋅ l
200
(18-13-2)
100 + α
W= ⋅w
200
Equ. 18-13-1 uses the individual fanning values, which depend on the ori-
entation of the corresponding anode face:
Fa: fanning between the anodes,

Fc: fanning between the anodes at the center of the pot,
Fl: fanning on that side of the anode which faces the long side of the pot.
Fs: fanning on that side of the anode which faces the short side of the pot.
ElysePrg calculates these Fanning Values with (see Fig. 18-13-4):
⎛ 0.1656 ⋅ IPD − 0.0043 ⋅ IPD3 + ⎞

Fi = ⎜ ⎟.
⎜⎜ 0.1270 ⋅ d − 0.0034 ⋅ d 2 + 0.0394 ⋅ IPD ⋅ d
⎟⎟
⎝ i i ⎠ i
(18-13-3)
⎡0.3844 + 0.006166 ⋅ h + ⎤
⎢ ⎥
( )
⎢⎣0.001822 ⋅ L + W − 0.000178 ⋅ h ⋅ L + W ⎥⎦ ( )
Figure 18-13-4 shows the meaning of the Parameter (distances, fanning

values, anode immersion depth, bath height) of Equ. 18-13-3.

18-71
Figure 18-13-4: Schematic representation of an electrolytic pot showing the meaning of

the paramters of Equ. 18-13-3.
Finally ElysePrg determines the Anodic Fanning Factor with:
Aeff
fA = (18-13-4)
Ageom
ElysePrg shows the calculation results (Fanning Factor, Effective Anode

Surface Area) and the parameters of Equ. 18-13-3 (distances, fanning
values, interpolar distance, bath height etc.) on the Anode Table Page
(Fig. 18-13-1).
Figure 18-13-4: Calculation Results and Paramters of Equ. 18-13-3 on the Anode Table
Page.

18-72
18.14. List Examples and Exercises Window
18.14.1. Overview
With the List Examples and Exercises Window you call ElysePrg to
execute standard calculation according to the examples and exer-
cises of ElyseSem.
You open the List of Examples and Exercises Window (Fig. 18-14-1) by
clicking on the corresponding item of the View Pulldown Menu.
Figure 18-14-1: List of Examples and Exercises Window.

This window contains the titles of all the examples and exercises in Elyse-
Sem. You click once on the line (for instance "Exs. 2-4 Alcoa Smelting
Process") and ElysePrg executes the example (Fig. 18-14-2)
Figure 18-14-2: Execution of exercise 2-4 with ElysePrg that was called from List Exam-
ples and Exercises Window.

18-73
The idea of the List of Examples and Exercises Window is to help you to
solve more easily the problems of everyday aluminum production. If you
have taken part in one of the seminars you may remember that an exam-
ple or exercise similar to his problem was solved during the seminar. If so,
you can search the List of Examples and Exercises Window for that ex-
ample/exercise, click on it and put your own values in the input boxes of
ElysePrg to get the solution of your particular problem.

18-74
18.15. Overview Plot Window
18.15.1. Overview
With the Overview Plot Window you determine selected bath proper-
ties for a the range of concentration of one electrolyte component.
ElysePrg plot these values on the Overview Plot Window.
Suppose you want to know how the properties of the electrolyte depend on
its composition. For instance you want to know how the bath and liquidus
temperature, the electric conductivity and the vapour pressure vary as the
alumina concentration increases from 0 to 10 %. You study this behaviour
with the help of the Overview Plot Window.
You open the Overview Plot Window (Fig. 18-15-1) by clicking on the
Overview Plot item of the View Pulldown Menu. Then you select the value
to be plotted on the x-axis, i.e. the alumina concentration and you define
the start (0 %) and end (10 %) concentration values. Finally you select the
values to be plotted on the y-axis, i.e. temperatures, electric conductivity
and vapour pressure.
Figure 18-15-1: Overview Plot Window showing a plot of the bath and liquidus tempera-
ture, the electric conductivity of the electrolyte and the vapour pressure
versus the alumina concentration.

18-75
After ElysePrg has plotted the curves of Fig. 18-15-1 you click somewhere
in the plot area, causing ElysePrg to plot the so-called Value Line. You
click then on the line and drag it to a different concentration value, 4% for
instance in Fig. 18-15-1. When you release the mouse button ElysePrg
updates all the other pages and windows using the new concentration
value. Fig. 18-15-2 shows as an example the Phys. Bath Properties Page
whose values Elyseprg has updated according to the position of the Value
Line in the Overview Plot Window.
Figure 18-15-2: Phys. Bath Properties Page after the user has placed the Value Line on
4 % Al2O3 in the Overview Plot (Figure 18-15-1). The page shows the
updated values corresponding to the new alumina concentration.
18.16. Literature
Lit. 18-1: M. W. Chase et al.,

"JANAF Thermochemical Tables Third Edition",
J. Phys. Cem. Reference Data Vol 14, 1985 Supplement 1.
Copies Only of Compounds Interesting for Aluminum Production.
Lit. 18-2: W. E. Haupin,

"Interpreting the Components of Cell Voltage",

18-76

Elyses em

Uploaded by

Copyright:

Available Formats

Elyses em

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Elyses em

Uploaded by

Copyright:

Available Formats

What is the purpose of the document?

What is the purpose of the document?

What topics are covered in the document?

What topics are covered in the document?

ALCAN ALESA ENGINEERING LTD.

Ulrich Heinzmann Practice of Aluminum Smelting.

Contents Electrolysis Seminar

Theory of Aluminum Smelting

Practice of Aluminum Smelting

17. How to Use ElyseSem

ALCAN ALESA ENGINEERING LTD.

ALCAN ALESA ENGINEERING LTD.

1.2. Aluminum Production by Electrolysis....................... 1-3

1.3. Industrial Production of Aluminum ........................... 1-4

1.4. Alternative Production Processes........................... 1-12

1.5. Inert Anodes .............................................................. 1-14

1.6. Wettable Cathodes .................................................... 1-18

1.7. Literature.................................................................... 1-21

ALCAN ALESA ENGINEERING LTD.

ALCAN ALESA ENGINEERING Ltd.

1. Principles of the Hall-Héroult Process

The principles of the Hall-Héroult process to produce aluminum

1.2. Aluminum Production by Electrolysis

In 1886 Hall of the USA and Héroult of France invented simultaneously

ALCAN ALESA ENGINEERING LTD.

In Figure 1-1 such a Hall-Héroult electrolysis cell is schematically shown.

1.3. Industrial Production of Aluminum

In Figure 1-2 you see a photograph of an industrial electrolysis cell

ALCAN ALESA ENGINEERING Ltd.

Figure 1-3 shows the corresponding schematic cross section of such a

Figure 1-3: Cross section of an industrial electrolysis cell.

ALCAN ALESA ENGINEERING LTD.

Pot Shell, Lining, Bath and Metal Pad

Electric Current Flow

ALCAN ALESA ENGINEERING Ltd.

Dense Phase Conveying System (Lit. 1-4)

Figure 1-4: Principle of the dense phase conveying system.

Hyperdense System (Figure 1-5)

Figure 1-5: Principle of the hyperdense system.

Point Feeder (Figure 1-6)

ALCAN ALESA ENGINEERING LTD.

Figure 1-6: Point feeder with metering device.

High Current Load Pots

ALCAN ALESA ENGINEERING Ltd.

Figure 1-7: The 200 kA Pot of Dubai Aluminium Co. Ltd.

ALCAN ALESA ENGINEERING LTD.

Figure 1-8: The 300 kA Pot of Pechiney.

ALCAN ALESA ENGINEERING Ltd.

Figure 1-9: The 300 kA Pot of VAW.

ALCAN ALESA ENGINEERING LTD.

1.4. Alternative Production Processes

Since the technical introduction of the classical Hall-Héroult-Process and

Direct Carbothermic Reduction Process

A crystallizer splits the alloy up into its components by fractional crystalli-

ALCAN ALESA ENGINEERING Ltd.

Alcoa Smelting Process

Al2O3 + 3 C + 3Cl2 = 2AlCl3 + 3 CO2 (1–1)

The AlCl3 is then decomposed to aluminum and chlorine by electrolysis:

2AlCl3 = 2Al + 3Cl2 (1–2)