Aluminium smelting

Aluminium smelting is the process of extracting aluminium from its oxide alumina, generally by the Hall-Héroult process. Alumina is extracted from the ore Bauxite by means of the Bayer process at an alumina refinery.

This is an electrolytic process, so an aluminium smelter uses prodigious amounts of electricity; they tend to be located very close to large power stations, often hydro-electric ones, and near ports since almost all of them use imported alumina.

Layout of an aluminium smelter

An aluminium smelter consists of a large number of pots, steel containers lined with carbon, in which the electrolysis takes place; smelting is run as a batch process, with the aluminium metal deposited at the bottom of the pots and periodically drained off. Power must be constantly available, since the pots have to be repaired at significant cost if the liquid metal solidifies.

The anodes are made of carbon, generally derived from pitch.

A typical smelter contains about 300 pots, each of which produces about a ton of aluminium a day, though the largest proposed smelters are up to five times that capacity.

Environmental issues of aluminium smelters

The process produces a quantity of fluoride waste: perfluorocarbons and hydrogen fluoride as gases, and sodium and aluminium fluorides and unused cryolite as particulates. This can be as small as 0.5 kg per ton of aluminium in the best plants in 2007, up to 4 kg per ton of aluminium in older designs in 1974. Unless carefully controlled, these fluorides tend to be very toxic to vegetation around the plants.

The Soderburgh process which bakes pitch to form the electrodes produces significant emissions of polycyclic aromatic hydrocarbons as the pitch is baked .

The linings of the pots end up contaminated with cyanide-forming materials; Alcoa has a process for converting spent linings into aluminium fluoride for reuse and synthetic sand usable for building purposes and inert waste..

Example aluminium smelters

Alcan Lynemouth Aluminium Smelter , powered by the coal-fired Lynemouth Power Station in North East England

Anglesey Aluminium , powered by Wylfa nuclear power station in north-west Wales (Shut down as of 30/09/09)

The Valco aluminium smelter in Ghana, powered by the Akosombo Dam hydro-electric dam

Fjarðaál in Iceland, to be powered by the Kárahnjúkar dam Jharsuguda in Orissa, India, to be powered by its own 1215 MW coal-fired

power station. Alcoa 's Point Henry smelter near Geelong, Australia, powered by its own

brown coal fueled power station and grid electricity.

Bauxite is the most important aluminium ore. It consists largely of the minerals gibbsite Al(OH)3, boehmite γ-AlO(OH), and diaspore α-AlO(OH), together with the iron oxides goethite and hematite, the clay mineral kaolinite and small amounts of anatase TiO2. It was named after the village Les Baux in southern France, where it was first discovered in 1821 by the geologist Pierre Berthier.

The Bayer process is the principal industrial means of refining bauxite to produce alumina.

Bauxite, the most important ore of aluminium, contains only 30-54% alumina, Al2O3, the rest being a mixture of silica, various iron oxides, and titanium dioxide [1] . The alumina must be purified before it can be refined to aluminium metal. In the Bayer process, bauxite is digested by washing with a hot solution of sodium hydroxide, NaOH, at 175 °C. This converts the alumina to aluminium hydroxide, Al(OH)3, which dissolves in the hydroxide solution according to the chemical equation:

Al2O3 + 2 OH − + 3 H2O → 2 [Al(OH)4]−

The other components of bauxite do not dissolve. The solution is clarified by filtering off the solid impurities. The mixture of solid impurities is called red mud, and presents a disposal problem. Next, the hydroxide solution is cooled, and the dissolved aluminium hydroxide precipitates as a white, fluffy solid. When then heated to 1050°C (calcined), the aluminium hydroxide decomposes to alumina, giving off water vapor in the process:

2 Al(OH)3 → Al2O3 + 3 H2O

A large amount of the alumina so produced is then subsequently smelted in the Hall-Héroult process in order to produce aluminium.

The Bayer process was invented in 1887 by Karl Bayer. Working in Saint Petersburg, Russia to develop a method for supplying alumina to the textile industry (it was used as a mordant in dyeing cotton), Bayer discovered in 1887 that the aluminium hydroxide that precipitated from alkaline solution was crystalline and could be easily filtered and washed, while that precipitated from acid medium by neutralization was gelatinous and difficult to wash.

A few years earlier, Louis Le Chatelier in France developed a method for making alumina by heating bauxite in sodium carbonate, Na2CO3, at 1200°C, leaching the sodium aluminate formed with water, then precipitating aluminium hydroxide by carbon dioxide, CO2, which was then filtered and dried. This process was abandoned in favor of the Bayer process.

The process began to gain importance in metallurgy together with the invention of the electrolytic aluminium process invented in 1886. Together with the cyanidation process invented in 1887, the Bayer process marks the birth of the modern field of hydrometallurgy

Today, the process is virtually unchanged and it produces nearly all the world's alumina supply as an intermediate in aluminium production

The Hall-Héroult process is the major industrial process for the production of aluminium. It involves dissolving alumina in molten cryolite, and electrolysing the solution to obtain pure aluminium metal.

Process

Aluminium cannot be produced by the electrolysis of an aluminium salt dissolved in water because of the high reactivity of aluminium. An alternative is the electrolysis of a molten aluminium compound.

In the Hall-Héroult process alumina, Al2O3 is dissolved in a carbon-lined bath of molten cryolite, Na3AlF6. Aluminium oxide has a melting point of over 2,000 °C (3,630 °F) while pure cryolite has a melting point of 1,012 °C (1,854 °F); a small percentage of aluminium oxide dissolved in cryolite has a melting point of about 1,000 °C (1,830 °F). Aluminium fluoride, AlF3 is also present to reduce the melting point of the cryolite.

The mixture is electrolyzed. This causes the liquid aluminium to be deposited at the cathode as a precipitate, while the oxygen from the alumina oxidizes the carbon anode to carbon dioxide. The electrical voltage across each cell is low (typically 3-5 volts DC), but a considerable amount of current is drawn by the circuit - in state of the art cells the cell current can be from 220 kA[1] to 340 kA.[2] The oxidation of the carbon anode reduces the voltage across each cell, increasing the electrical efficiency at a cost of releasing carbon dioxide into the environment. Hundreds of Hall-Heroult cells are usually arranged in series and supplied from a single transformer set that generates the current with a voltage of 1 - 2 kV from 110 kV or more high voltage supply lines. The heavy current is supplied through heavy busbars usually made of cast aluminum. The cells are electrically heated to reach the operating temperature with this current, and the anode regulator system varies the current passing through the cell by raising or lowering the anodes and changing the cell resistance. If needed any cell can be bypassed by shunt busbars.

Hall-Heroult Industrial Cell/Pot

The liquid aluminium is taken out with the help of a siphon operating with a vacuum to avoid having to use high temperature valves and pumps. The liquid aluminium then may be transferred in batches or via a continuous hot flow line to the casting facilities. The metal is then either cast into the final forms with any alloying materials needed, or cast into ingots that are remelted.

While solid cryolite is denser than solid aluminium at room temperature, the liquid aluminium product is denser than the molten cryolite, at about 1000 °C, and sinks to the bottom of the bath, where it is periodically collected.[3] The top and sides of the bath are covered with a crust of solid cryolite which acts as thermal insulation. Electrical resistance within the bath provides sufficient heat to keep the cryolite molten.

As the proportion of alumina is depleted in the cryolite additional alumina is added by a hopper system to maintain the alumina composition. The solid crust at the top of the bath prevents this and the crust is periodically broken to allow the added alumina to mix in with the electrolyte.

The electrolysis process produces exhaust which escapes into the fume hood and is evacuated. The exhaust is primarily CO2 produced from the anode consumption and hydrogen fluoride (HF) from the cryolite & flux. HF is a highly corrosive gas and attacks glass surfaces which means that cranes and heavy equipment used in the plant need glass windscreens and windows to be covered with plastic film. The gases are usually treated in adjacent treatment plants which dissolve the HF in water and neutralize it. The particulates are also captured and reused using electrostatic or bag filters. The remaining CO2 is exhausted into the atmosphere.

The large current causes heavy magnetic fields, and can stir the aluminium with magnetohydrodynamic (MHD) forces. The stirring of aluminium in the cell increases its performance, but reduces the purity, since the materials get evenly mixed. Otherwise the cell can be operated with static aluminium pool so that the impurities either rise to the top of the aluminium or sink to the bottom leaving high purity aluminium in the middle.

Aluminium smelters are usually sited where economical hydroelectric, wind or fossil fuel power is available. In some european smelters, the electrical energy produced in countries such as Norway is transported via high voltage lines to Germany and other areas and used by smelters. Since aluminium smelters require constant supply they allow the best use of constant generation capacity, and can be used to increase the base load to make the demand more constant and less cyclical. This can make the overall electrical generation and transmission system more economical for end users.

The need of electrical power and pollution of the surroundings were early problems with this reaction. The use of hydroelectric power plants and new filter systems has resolved this to some extent, but the problem still exists.[citation needed]

History

The Hall-Héroult process was discovered independently and almost simultaneously in 1886 by the American chemist Charles Martin Hall [4] and the Frenchman Paul Héroult. In 1888, Hall opened the first large-scale aluminium production plant in Pittsburgh, which would eventually evolve into the Alcoa corporation.

In 1997 the Hall-Héroult process was designated an ACS National Historical Chemical Landmark in recognition of the importance of the commercialization of aluminium.

Aluminum, an overview

Aluminum is the most abundant metallic element in the Earth’s crust (about 8%) and is the third most common element after oxygen and silicon. Unlike copper or gold, aluminum cannot be found in nature in the pure state because of its high affinity with oxygen, being so always combined with another element like in alum (KAl(SO4)2∙12H2O) and in aluminum oxide (Al2O3). So, up to 1820, the aluminum was unknown as a metal.

In the 19th century the production process was so expensive and available quantities so small that aluminum was a precious metal ($1200/kg in 1852). Indeed, Napoleon III emperor of France had a baby rattle and some other small objects made of it, and a story tells that during a banquet the most honoured guests were given aluminum utensils, while the other guests were given gold utensils.

Presently, aluminum is the second largest used metal in the world, mainly due to its light weight, high strength and recyclability.

Aluminum is heavily used in the transportation industry because of its durability, strength and lightweight. Aluminum weight is one third of steel or cast iron. Taking into account increased thickness of the aluminum parts compared to steel, 1 kg of aluminum replaces 2 kg of steel, leading to lighter cars, trucks, etc… with reduced fuel consumption and CO2 generation.

Without aluminum the commercial aircraft industry would not have existed. The new A380 employs 66% of aluminum in the airframe, while a Boeing 747 contains 75 tons of aluminum.

The use of aluminum for the building of ships is increasing year by year. Today, single and multiple hull boats are made entirely of aluminum alloy. This kind of marine applications involve the largest usage of aluminum per produced object (400 ton) compared to a large, all aluminum car (1 ton).

The usage of aluminum is increasing in the military field too, where it is used as a substitute for the steel.

In building and construction aluminum find a wide variety of applications, and its use is steadily increasing. It can be used to manufacture structural elements, as in bridges (for example, the Corbin Bridge in Pennsylvania has been retrofitted with an extruded aluminum deck, which is lighter than the previous deck made of steel and timber, allowing the bridge to sustain 22 tons load compared to the previous 7 tons). Curtainwall made of extruded aluminum and glass are very attractive for the design of new buildings or retrofit of old ones. Windows made of extruded aluminum are attractive, energy-efficient (with thermal broken technology), and reliable. Domes for gymnasiums, schools, theme parks, storage facilities, multi-purpose arenas, industrial roof systems, and churches are made with aluminum because of its strength and low weight. Aluminum is one of the best material also in the roof construction, because of its strength against corrosion and, hence, weathering and influence of pollutants in the atmosphere. Low maintenance aluminum facades are used to cover old houses facades made with thin or wide wooden clapboard.

Aluminum find also wide use in the packaging industry, being produced in both rigid and foil forms. Rigid aluminum containers are used for beverage and food packaging. Aluminum cans account for all of the beverage can market, but only a small percentage of the food can market. Cans are 79 percent of aluminum packaging by weight. Foil packaging is used as a wrapping foil, as semi-rigid packages such as pie plates and frozen food trays, and as flexible packaging such as cigarette foil and candy wrappers.

Aluminum history

The term aluminum comes from the alum compound, (KAl(SO4)2∙12H2O), which contains it. In 18th century it was realized that alum contained a peculiar base, different from all others, but only in the 19th century scientists were able to isolate it.

Berzelius and Davy tried to use electrolysis processes to isolate the metal contained in alumina. In particular Davy heated a mixture of alumina and potash and then submitted the mix to a electrical current, but with no success.

In 1825 was Oersted to try to isolate the metal believed to be contained into the alumina. He first prepared anhydrous aluminum chloride (AlCl3) by passing a current of chlorine over a mixture of charcoal and alumina heated to redness. Then, the aluminum chloride produced was mixed and heated with potassium amalgam producing potassium chloride and aluminum amalgam. Distillation of this amalgam without contact with the atmosphere left a lump of metal which color, said Oersted, resembles tin in color and sheen.

Wöhler repeated the Oersted experiment, but was unable to duplicate the results. Recent repetitions of the experiment, based on Oersted notes, have shown that using a very dilute potassium amalgam (1.5% K) it is possible to extract the aluminum.

Oersted encouraged Wöhler to continue in his attempts. So, Wöhler first produced the anhydrous aluminum chloride. Then, he put in a platinum crucible a mixture of AlCl3 and potassium (K), instead of using potassium amalgam as Oersted. The crucible was closed and some heat applied just to start the reaction. Soon, the crucible became very hot, being the reaction:

exothermic. The produced aluminum was contaminated with platinum, so he repeated the experiment with porcelain and other materials crucibles, always being able to produce aluminum in the form of a gray powder. The aluminum produced was contaminated by potassium, platinum or aluminum chloride, however, he first described the most important properties of aluminum and then, in 1845, was able to produce a coherent mass of aluminum.

In 1855 French scientist Henri Sainte-Claire Deville, ignoring Wöhler 1845 experiment, duplicated it. Observing the aluminum chloride reduction, he understood the importance of the discovery and immediately started to study how to produce large quantities of this metal to be commercialized. Two ways were possible. The first one was the chemical production of the aluminum, using the chloride as salt and the potassium as the agent to reduce it, the second choice was an electrolytic method to reduce the chloride, with a carbon anode and a platinum cathode. Both methods were difficult to implement on a large scale. In the chemical method, the potassium was very expensive and dangerous to manipulate, while there were not available sources of electricity at low cost, the only one existing being primary batteries. So, H. St.Claire Devile choose the chemical method to produce aluminum, using as reduction agent the most common in nature sodium (Na) rather than potassium. The

Deville’s work was greatly helped by the funds gave him by Napoleon III. Thanks to the work of other chemists of that time, Deville was able to produce for himself large quantities of sodium. At the same time in this period two other processes had to be converted into industrial processes, the first one being the production and refining of alumina and the second one the conversion of alumina into aluminum chloride. Anyway, in 1859 the available quantity of aluminum increased (2 tons/year produced in Salindres plant). Through the years going from 1855 to 1890 the aluminum price declined from the equivalent 1200 €/kg to 13 €/kg. During that years, different chemical and electrochemical methods were developed for the aluminum extraction. Chemical methods were more or less variations around the basic St. Claire Deville process. We can mention: Castner, Netto, Grabau, Webster, Frismuth, etc…The electrochemical processes could be splitted into two main categories: electrothermic and electrolytic. For the electrothermic processes we can mention Cowles, Héroult, Brin, Bessemer, Stefanite, Moissan, while for the electrolytic processes we have Héroult, Adolphe Minet, Hall, Hampes, Kleiner, Gooch, Waldo.

From all these processes, emerged the work of the American Charles Martin Hall and French Paul Louis Toussaint Héroult, whom independently and at the same time invented the process that is still used to reduce alumina to aluminum.

In 1880 at the Oberlin College near Cleveland, Ohio, two people met. One was Frank Fanning Jewett, professor of chemistry and mineralogy at the Oberlin College. Jewett had the possibility to study at the University of Göttingen in Germany between 1873 and 1875. During his stay in Germany he met Wöhler, who had isolated aluminum in 1827. The other was a young student from Oberlin, Charles Martin Hall, self-educated person in chemistry and so passionate to conduct experiments at his home. Hall, at the age of 16, already knew the problems involved in the aluminum production and, under the Jewett guidance and encouragement, worked on aluminum chemistry both at home and Oberlin College laboratory. Hall first tried chemical reduction methods to produce aluminum, but with no results. Then, together with Jewett, they decided to follow the electrolytic way to produce aluminum. In a first attempt, Hall tried to decompose aluminum fluoride (AlF3) dissolved in water, but the only results was the production of hydrogen at the anode and aluminum hydroxide at the cathode. The important thing in this work was the selection of the fluoride, never tried before. The next step was the use of fused salts as solvents. After different experiment, he choose the cryolite as the solvent, being him and Jewett aware of its use in the Deville works. He was able to melt cryolite and showed that it was a good solvent for alumina. The first electrolysis attempt was made using a clay crucible, but the result was the production of silicon from the crucible material. Then, recognizing this, the clay crucible was lined with graphite, it was added some aluminum fluoride to the cryolite in order to lower the melting point and the experiment was repeated. After several hours of electrolysis, he cooled the melt and broke it open, finding small silvery globules that Jewett confirmed to be aluminum. It was February, 23rd 1886.

After some troubles, in 1888 a group of investors, organized by Captain Alfred Hunt, provided Hall with enough fund to establish the Pittsburgh Reduction Company, the Alcoa predecessor.

One or two months later the Hall discovery, French Paul Louis Toussaint Héroult made the same discovery independently. He learned of the St.Claire Deville

discoveries at the age of 15 reading the Deville’s famous treatise. He tried to produce aluminum electrolyzing various aluminum compound, with no success. Then, he tried to electrolyze cryolite. During the process, he find the iron cathode was melted. Because the temperatures were not so high to justify a melting of the iron, he realized that some alloy had been formed. Few days later, in an attempt to lower melt temperature he added some sodium aluminum chloride and noticed that the carbon anode was consumed. What happened was that the sodium aluminum chloride entered in contact with the moisture, being converted into hydrated alumina. So, basically Héroult discovered the same process to produce aluminum revealed by Hall just a couple of months before.

Process basics

The aluminum is produced extracting it from the aluminum oxide (Al2O3), called also alumina, through an electrolysis process driven by electrical current. The process uses as electrolyte a molten salts called Cryolite (Na3AlF6) capable of dissolve the alumina. Carbon anodes are immersed into the electrolyte (usually referred as the "bath") carrying electrical current which then flows into the molten cryolite containing dissolved alumina. As a result, the chemical bond between aluminum and oxygen in the alumina is broken, the aluminum is deposited in the bottom of the cell, where a molten aluminum deposit is found, while the oxygen reacts with the carbon of the anodes producing carbon dioxide (CO2) bubbles. The alumina reduction process is described by the following reaction:

Once passed through the bath, the electrical current flows into the molten aluminum deposit and is then collected by the bottom of the pot, usually called "cathode".

The following is a schematic picture of an aluminum electrolysis cell:

Prebake and Soderberg

The alumina reduction cells, and in a broader sense, the aluminum smelters can be divided into two big categories depending on how is arranged their anodic system.

In the alumina reduction cells the anode is a block of carbon made of petroleum coke and pitch. What differentiate the two technologies is the way this carbon block is produced.

In a so-called pre-bake cell, the petroleum coke is mixed with pitch, which acts as a binder. Then, at this mixture, usually called green paste, it is given a parallelpiped shape with either a press or a vibrocompactor. The formed carbon block is then baked into furnaces in order to be transformed into a solid carbon block. The electric current arrives to the carbon block through a rod linked to it through nippels. A pre-bake pot contains several single anodes (usually 14 ÷ 40, mainly depending on the line current), which stay on the pots for a fixed amount of days (generally 26 to 30 days). Then, before being completely consumed, they are removed together with the rod, and the remaining carbon reused to produce new anodes.

In a Soderberg smelter the basic idea is to eliminate the sub-plants which form, bake and join the carbon block with the rod. A Soderberg cell has only one big anode, housed in a steel container, which gives to the anode its shape. From the upper part of this container it is introduced the green paste. During its movement from the top to the bottom of the container the green paste is baked. Unfortunately, the quality of the baked Soderberg anode is lower than the quality of the prebaked one, hence the Soderberg cells are always characterized by a lower current efficiency and a higher pot voltage, needed also to produce the extra heat necessary for the anode baking.

Presently all the new built smelters adopt the pre-bake technology, because of the higher current efficiency, lower specific energy consumption and lower emission (especially PAH). However, a good number of Soderberg plants are still in operations, sometimes retrofitted with additional technology aimed at increasing current efficiency and reduce emissions.

Throughout the rest of this website we will refer only to the pre-bake technology, even though most of the topics apply both at pre-bake and Soderberg technology.

Detailed Description of a Cell and its Basic Functioning

Even though the basic elements in the aluminum production are still the same invented by Hall and Héroult more than 100 years ago, a modern industrial electrolysis cell is different from the first pots of the 19th century.

First of all, the size of the cells is changed to accommodate for the much larger current intensities used nowadays. Secondly, in a modern pot we find a series of elements used to reduce the energy consumption and the gas emissions.

The alumina reduction occurs in a vessel which is made of several parts designed altogether to:

Act as a container for the molten bath and aluminum

Resist to the high temperatures (around 950°C) of the molten liquids it contains

Resist to chemical attacks brought especially by the molten electrolyte constituents

Resist to wearing caused by alumina abrasive behavior

Reduce heat losses to a technical and economical optimal minimum

Be mechanically enough resistant, but also with sufficient elasticity in order to

accommodate for the thermal and physical expansion of the materials it contains

Collect the electrical current coming from the anodes with a minimum voltage drop

To achieve all these properties the vessel is made with a combination of different materials. First, we find an outer steel container, usually referred as potshell, which contains

all the other elements of the cell. On the potshell bottom are deposited some layers of thermally insulating bricks with the aim of reducing heat losses from the bath. Above this, we find a refractory bricks layer which are very resistant to prolonged periods of exposition to the cell high temperatures. The carbon blocks that physically constitute the container of the molten aluminum and electrolyte stay above these bricks layers. The bottom blocks are called cathodes, even though electrochemically speaking is the metal pool which acts as cathode. They also collect the current exiting from the metal pad. The container sides are made with other carbon blocks. Cathodes and side blocks are joined together with a mix of pitch and carbon dust that is pressed inside the joints between them. At the bottom of every cathode we find a slot that accommodates an iron bar, called collector bar, with the purpose to transport outside of the cell the current collected by the cathode. Cathode and collector bar are joined together filling the space between them with molten cast iron, which subsequently freezes bonding together the parts.

The vessel described above contains, as said before, the molten aluminum and electrolyte. Due to the different densities, the molten bath stays above the molten metal, as the oil stays on the top of the water to give a practical example. In the molten bath occur all the chemical reactions for the alumina reduction. This reactions are driven by the electrical current transported inside the bath by the carbon anodes partially immersed into the molten bath.

The term “anode” refers not only to the carbon block. More in general an anode is made by a rod, a yoke and a series of stubs (1 to 6 generally) partially housed in rounded cavities obtained in the carbon blocks. Anode rods are made with copper or aluminum while yoke and stubs are made with iron. The carbon part of the anode is joined together with the metallic part of the anode assembly pouring molten cast iron in the space between the anode cavity and the rod. The molten cast iron freezes joining together the carbon block and the metallic part. In a modern pot we can find up to 40 anodes and because the carbon of the anodes participates to the chemical reactions, hence being consumed, they need to be replaced on a regular basis.

All the anodes of a cell are fixed to an aluminum structure, called “bridge”. The bridge transports the electrical current to the anodes and is also equipped with an electrical motor and a series of levers in order to raise or lower all the anodes of a cell. In this way it is possible to control the voltage at which the pot is operating.

Because of the high working temperatures of the anodes, on the top of them is put a protective layer of material made of a mix of crushed bath and alumina, in order to avoid their burning to the air. This protective layer is called “crust”.

The CO2 formed by the reaction of the anodes with the alumina oxygen escapes from the bath as gas bubbles, while the aluminum, being no more bonded with the oxygen is deposited in the metal pad inventory, increasing its height as the production goes on.

As the aluminum production proceeds, the alumina dissolved into the bath is depleted, and needs to be restored on a regular basis. Each cell, hence, is equipped with an alumina bin with a feeding system which delivers alumina to the electrolyte.

This feeding system is made of a crust breaker, basically made with a steel rod operated with a pneumatic cylinder, which opens a hole on the crust, and a corresponding alumina feeder, which dumps a certain amount of alumina into the molten bath from the hole opened in the crust by the crust breaker. The bath is then restored in its alumina content. Depending on the pot size, each pot can have 1 to 6 crust breakers and alumina feeders. On some pot technologies, crust breaker and alumina feeder are integrated.

The feeding operations are usually controlled by a computer control system following some algorithms.

To collect the fumes escaping from the pot crust, due to the bath evaporation, the pots are completely closed with removable pot covers, which collect the gases coming from the bath and direct them towards the gas treatment center.

How an aluminum smelter is made

In an aluminum smelter usually we find several hundred pot, housed in one or more large buildings called potrooms. Inside the potroom the cells are electrically connected in series. This means that the cathode of a cell is connected with the anodes of the next cell downstream. The electrical connection of one cell to the next one is made with a complicated system of aluminum bars with cross sectional areas suitable to transport the so high currents used with minimum voltage drops. In the most recent smelters the cells are arranged in a side by side configuration, which makes easier to lessen the adverse effects of the magnetic fields. In older smelters it’s easy to find also the end-to-end configuration.

As we have seen, a cell for its basic functioning needs alumina, carbon anodes and electrical current. In an aluminum smelter, hence, together with the potroom, we find other buildings and sub plants which deliver to the potroom carbon anodes, alumina and electrical current.

The carbon plant produces the anodes for the potroom. Anodes are made starting from petroleum coke, pitch and recycled butts. The coke is grinded in various fractions and then is mixed together with crushed butts and pitch. This mix is then heated and vigorously stirred in a plant called green mill. The green paste produced is then transferred into a vibrocompactor or a press in order to produce a so called green anode. Green anodes are then sent to the baking furnaces, where they are heated up and stay for some time at around 1100°C, until they form a solid block of carbon. The baked anode is then sent to the rodding room where it is coupled to the anode rod with cast iron. The rodded anodes are sent to the potroom generally through a conveyor. The rodding room also receives the spent anodes from the potroom. These spent anodes are cooled down, the remaining cover material is removed from the anode and the remaining part of the carbon is stripped from

the rod. The carbon is then transported into the green mill where it will be reused to produce other new anodes. In the carbon plant are present also treatment systems for the gases produced by the baking of the anodes.

Alumina is transferred from the main smelter silos to the potroom usually by conveyors or air slides. Before entering into the potroom, the alumina is used to “clean” the gases evolving from the crust of the pot. The cleaning is realized finely mixing the fresh alumina with the hot gases escaping from the cells rich in fluoride. The alumina adsorbs on its surface the fluoride, which then goes back to the potroom together with the alumina.

The alternate current received from the external grid needs to be converted into a low voltage, high amperage direct current. This is done by a dedicated plant, which lowers the voltage with transformers and then rectifies the current using a series of diodes in parallel.

After the potroom it is placed the casthouse, which receives the molten metal produced by the cells and after some processes (skimmimg, degassing, alloying) produce the end products as ingots, billets or slabs

Process thermodynamic - Enthalpy

As we have seen, in a Hall-Héroult cell the molten aluminum is produced according to the following reaction:

This reaction occurs at constant temperature T (the bath temperature) and constant pressure p (the atmospheric pressure).

Enthalpy

When we speak of enthalpy, we are talking of the 1st law of thermodynamic or, in other words, of the energy conservation principle. In the case of an alumina reduction cell, from the external the only energy we input is electrical energy (WEl). This energy is partially used by the cell to produce aluminum while the rest is dissipated as heat. So we can write:

This equation tells us that the electrical energy we supply from the external is used to: 1) increment the internal energy content of the products of the reaction (1) compared to the internal energy of the reactants (term ΔU), 2) give the energy necessary for the CO2 bubbles to expand under the atmospheric pressure p (term pΔV), 3) the remaining part of W El is lost in the external as heat (term QD).

We also know that the term (ΔU + pΔV ) is equal to the enthalpy change ΔH, so we can rewrite the (2) as:

Using thermodynamic tables and the basic reaction (1) we can calculate the enthalpy of the reaction, or, in other terms, the minimum energy required to produce aluminum.

But for a better estimation of the enthalpy we need to take in considerations the following practical aspects:

We need to consider the energy necessary to heat the reactants from the ambient

temperature up to the bath temperature

The process is not 100% efficient, so we need to take into account the loss in current

efficiency

With these considerations, if x is the current efficiency expressed as a fraction (CE = x ∙ 100%), (1-x) is the fraction of aluminum which reoxidizes and the (1) can be rewritten as:

Considering an electrolyte saturated with α alumina, with PCO2 = 1 atm, at a temperature of 977°C, inserting the proper figures from the thermodynamic tables we have:

This equation takes into account:

The energy required to heat the reactants from ambient temperature to 977°C

A current efficiency lower than 100%

At 100% current efficiency:

The (5) should be further modified if we consider an electrolyte not saturated with alumina (as is the real case) and the fact that usually the alumina delivered to the cell is mostly γ alumina, but these changes introduce only slight modifications to the (5).

Process thermodynamic - Free energy

Gibbs Energy

To better explain the Gibbs energy concept applied to the alumina reduction, we will use first as an example the process of electrolysis of the water, depicted in the picture below, analyzed in condition of reversibility:

The total energy necessary to electrolyze the water is equal to ΔH = 285.83 kJ/mol. This amount includes:

The change in the internal energy U linked with the change in the kinetic, potential

and chemical bonds energy of the species involved in the electrochemical reaction

The energy required by the gas bubbles (O2 and H2) formed during electrolysis to

expand under the pressure p:   work

being, as we know:

Now arise a question: do we have, with the battery, to provide all the energy linked to the change in enthalpy? Let’s see…

The change in the internal energy U is equal to:

Where L is the work applied to the system while Q is the heat given to the system. In the case of a reversible reaction at constant temperature T the term Q is equal to TΔS. Furthermore, in our case the work L has to parts:

The electric work inputted by the battery: WEl

The work done by the gases expanding: -pΔV

Hence we have:

Substituting the (8) into the (6):

Now, in our example, the only kind of external energy we are deliberately providing to the water is the electrical energy of the battery, and this energy is equal to:

Since the reaction is characterized by an increase in entropy (ΔS > 0), the energy W El that the battery has to provide from the external is less than the total energy required by the electrolysis reaction. The “missing” energy is taken as heat from the environment.

The term ΔH-TΔS is equal to the Gibbs energy, or “free” energy. So we can write:

Let’s take again in consideration the (1), the basic alumina reduction reaction, in reversible conditions (hence at 100% current efficiency), occurring in an environment at temperature T. For this reaction remain valid all the arguments we have developed for the previous example. Let’s calculate now the amount of electrical energy needed by the (1) to occur.

If we take again in considerations the reaction (4), with the same assumptions (electrolyte saturated with α alumina, PCO2 = 1 atm, temperature of 977°C), inserting the proper figures from the thermodynamic tables we have:

At 100% current efficiency we have:

The difference:

Is equal to the heat that is taken by the reaction (1) from the external environment. In the case of an alumina reduction cell the external environment is the bath itself, so the bath at its high temperatures (around 950°C) gives heat to the reaction (1). Obviously, the bath can not give indefinitely heat to the reaction, but this heat needs to be restored from the external. Practically, this means that we need to spend electrical energy also for the term TΔS.

Now we will calculate the minimum voltage to apply to a cell for the reaction (1) to occur in reversible conditions. In the (1) 12 moles of electrons are involved to produce 4 moles of aluminum.

In other terms, if F is the Faraday constant (the amount of electrical charge carried out by a mole of electrons) and E is the electrical potential we have:

Inserting the values we get:

Depending on the temperature at which occur the reaction (940°C ÷ 960°C normal operating ranges in the industry). This voltage is equivalent to the minimum energy, in reversible conditions, to give from the outside for the reaction (1) to occur. The voltage needed to compensate for the term TΔS is equal to:

The Voltage Drop in the Electrolyte

The picture below depicts an anode immersed into the electrolyte, the average distance of the anode bottom from the metal pad, called ACD, Anode to Cathode Distance, the thickness of the CO2 bubbles layer underneath the anode, indicated with the greek letter and the thickness of the layer of bubbles adhering to the anode, a.

This layer of bubbles is very important because it affects several aspects of the pot performance like:

Current Efficiency

Voltage

Noise

Talking about the voltage drop in the electrolyte, it is fundamental to point out that the current does not flow inside the CO2 bubbles. This implies that in calculating the voltage drop in the electrolyte we need to consider the actual areas available for the current to flow between the anode and the metal pad.

As depicted above, we have a layer of bubbles adhering to the anodes with a thickness a, while the thickness of the whole bubble layer is equal to .

Therefore, using the first and the second Ohm’s law and indicating the electrolyte conductivity with χ, the electrical resistivity ρ=1/ χ and the line current as I, we have:

Where:

Aa is the total area of the anodes, i.e. the area of a single anode multiplied by the total

number of anodes present in the pot

Aab is the total bottom area of the anodes free from bubbles adhering to them,

considering all the anodes present in the pot

Afb is the total average cross-sectional area of electrolyte free from bubbles,

considering all the anodes present in the pot

In some papers the amount of anode area covered by bubbles, Aab, adhering to them and the cross sectional area available for the current to flow in the bubble layer, Afb, is expressed as a function of the alumina concentration in the electrolyte.

More in general, the dynamic of the CO2 gas bubbles escaping from underneath the anode is very complicated and depends not only from the electrolyte composition but also from the single anodes dimension and spacing of one to the other.

From the equation above, we can understand how, with the same ACD, the presence of the bubbles layer increases the Vb voltage component through the decrement of the actual area available for the current to flow into the electrolyte.

This is the reason why in recent years most companies have developed the so called “slotted anodes” technology with the aim to reduce the bubbles layer thickness, hence reducing the VB voltage component or, working with the same VB, increase the ACD to improve the current efficiency.

Theoretical Production of Aluminum

Let’s consider the alumina reduction reaction:

This reaction involves the flow of 6 moles of electrons from the cathode to the anode to

produce 2 moles of aluminum.

This means that, to produce 2 moles of aluminum atoms, which weight:

are involved 6 moles of electrons, hence 6 Faradays of electrical charge:

Combining the (1) and (2) for each Coulomb of electrical charge flown into the electrolyte are produced:

So, in 24 hours a cell with a line current of I amperes produces:

Some Important Figures

Current Efficiency

From the electrochemistry laws we know that, if I is the current (in kA) flowing into a cell, the production of aluminum in 24h is equal to:

Theoretical Amount [in kg] = 8,0533 * I

For example, for a pot with a 200 kA flowing current the daily production should be equal to:

Theoretical Amount = 8,0537 * 200 = 1610.74 kg

But the actual weight of the aluminum produced is always lower than the theoretical amount. This is because together with the alumina reduction reaction:

Is always present a parasite reaction, called “back reaction”:

which reoxidizes to the state of alumina part of the aluminum produced.

The ratio:

Is called current efficiency and is always less than 100%. The best in class aluminum smelters operate with current efficiencies of 95 – 96%, while normal figures for current efficiencies range between 90 ÷ 94%.

Specific Energy Consumption

For the current to flow we need to apply a voltage V to the pot. The specific energy consumption, expressed in kWh/kg Al produced, from the electrochemistry laws is equal to:

This equation tell us that to reduce the specific consumption we need either to reduce voltage and/or increase current efficiency. In the aluminum industry specific energy consumption ranges between 13 ÷ 15 kWh/kg Al.

The theoretical amount of energy needed to produce aluminum at 100% current efficiency is equal to 6.34 kWh/kg. The difference between this theoretical and the actual amount in the real world is due to the fact that:

Current efficiency is never 100%

More than 50% of the electrical power given is lost as heat escaping from the pot

itself

Specific Carbon Consumption

As we have seen, in the alumina reduction it is involved also the carbon. Theoretically, at 100% current efficiency, to produce 1 kg of aluminum 0,333 kg of carbon are needed.

The real consumption is higher than the theoretical because current efficiency is always less than 100% and also because of the oxidation of the anodes to the air.

Practically, the specific carbon consumption is equal to:

Bath Chemistry

Criolyte Ratio, Mass Ratio and % Excess AlF3

Usually, in a modern point feed pot the bath composition is the following:

AlF3: 10% ÷ 13%

Al2O3: 2.5% ÷ 3.5%

CaF2: 4% ÷ 7%

MgF2: 0% ÷ 1%

The rest is cryolite

Calcium fluoride and magnesium fluoride typically build up into the bath because they arrive as alumina impurities.

Al2O3 is the raw material from where the aluminum is produced, while AlF 3 is intentionally added to increase the current efficiency.

Cryolite can be also thought as composed by three moles of sodium fluoride (NaF) and one mole of aluminum fluoride (AlF3):

Consequently, in the bath is always present some AlF3, coming from the cryolite. In the industrial process in addition to this AlF3 it is added some other AlF3. There are several ways to measure the bath composition respect to the NaF and AlF3 content:

Cryolite ratio is the ratio between:

For the pure cryolite the cryolite ratio is equal to 3

The weight ratio is equal to:

So, for the pure cryolite we have:

1 NaF mole weights: 1 Na mole (22.9897 g) + 1 F mole (18.9984 g) = 41.9881 g

AlF3 mole weights: 1 Al mole (26.9815 g) + 3 F mole (3 18.9984 g) = 83.9767 g∙

The AlF3 added can be expressed also as % excess AlF3, where the

excess is respect the AlF3 coming from the cryolite. In other terms:

Electrolyte Ionic Structure

Molten cryolite is completely ionized according to the following reaction:

Some

further ionizes as:

The AlF3 added in excess reacts with the F- ions in order to form AlF4- ions according to the

following reaction:

The AlF3 added acts as a Lewis acid accepting a lone pair (). This is the reason why the percent of excess aluminum fluoride is also referred as “acidity of the bath”.

When we add alumina to the cryolite bath this dissolves forming the following species:

Electrode Reactions

The reactions that guide the aluminum production are the following:

At the cathode:

Where the electrons come out from the metal pad. The formed aluminum is then deposited into the molten aluminum pad.

At the anode thermodynamically should be produced

CO, but actually is CO2 what evolves, because the reaction producing

CO2 is kinetically favoured compared to the reaction producing CO:

Electrolyte properties

The electrolyte used to dissolve the alumina has several properties that change

according to its chemical composition. Discussing on the various electrolyte properties we will see that the same changes in bath composition will improve some properties while worsening some others. As a result, there is not an optimum bath composition, but it is always fundamental to look for an optimal bath composition, which will be the best compromise between the various electrolyte properties.

The main electrolyte properties will be reviewed with a brief discussion on the impact on the process.

Liquidus Temperature

The electrolyte liquidus temperature, i.e. the temperature at which the electrolyte starts to freeze, depends on its composition and plays a very important role on many aspect of the process, including the heat balance and alumina dissolution.

Generally, we can say that every additive lowers the liquidus temperature. In particular, the following table gives only qualitative informations about the effect of the various additives on the liquidus temperature:

Increased Bath Component Qualitative Changes in Liquidus TempCaF2 ↓AlF3 ↘, not linearlyLiF ↓

MgF2 ↓NaCl ↓NaF ↘, not linearly

Al2O3 ↓Temp -

Alumina solubility

The electrolyte has the ability to dissolve in it the alumina, but up to a certain percentage. Above this percentage, the alumina added to the bath will be deposited into the bottom of the pot, disturbing the current flow and generating metal pad motion with decreased current efficiency as a result.

The alumina saturation concentration depends on the electrolyte composition, but generally we can say that every additive (except for KF, rarely used) lower the alumina solubility. Also the electrolyte temperature has an effect on the saturation concentration.

Again, the following table will give only qualitative information on the impact of the various additives on the alumina solubility:

 

Increased Bath Component Qualitative Changes in Al2O3 SolubilityCaF2 ↓AlF3 ↓LiF ↓

MgF2 ↓NaCl ↓NaF ↗↘, first up, then down

Al2O3 -Temp ↑

The following chart shows quantitatively the effect of the temperature on the alumina solubility, with a given bath composition:

Electrical Conductivity

The electrical conductivity of the bath has a direct impact on the pot voltage. In fact, working with a constant ACD, if the conductivity of the bath increases the pot voltage will decrease, or if we want to keep constant the pot voltage, an increase in the bath conductivity will increase the ACD.

Qualitatively, the effect of the various bath additives on electrical conductivity is:

Increased Bath Component Qualitative Changes in Electrical ConductivityCaF2 ↓AlF3 ↓LiF ↑

MgF2 ↓NaCl ↑NaF ↑

Al2O3 ↓Temp ↑

Density

The electrolyte floats above the metal pad because between bath and molten aluminum there is a density difference. The bath is “lighter” than the molten aluminum, so it floats above the molten aluminum like the oil above the water.

To achieve a good current efficiency it is fundamental that the two liquids, the bath and the aluminum, are clearly separated, and this is achieved with a electrolyte composition that gives the maximum difference in the densities between the bath and the aluminum pad.

The density of the molten aluminum is around 2.38 gr/cm3 while the bath density ranges between 2.07 and 2.15 gr/cm3. Again, the following table gives qualitative informations on the effect of the various bath additives:

Increased Bath Component Qualitative Changes in Bath DensityCaF2 ↑AlF3 ↓LiF ↓

MgF2 ↑NaCl ↓NaF ↗↘, first up, then down

Al2O3 ↓Temp ↓

Viscosity

The bath viscosity affects the hydrodynamic processes of a pot, where the electrolyte has a very complex pattern of movements due to gas bubble release, metal pad movement, etc…

Generally speaking, an increase in the bath viscosity will decrease the diffusion rate of the aluminum from the metal pad to the bath, hence increasing the current efficiency.

The qualitative viscosity behavior is shown in the following table as a function of the bath additives:

Increased Bath Component Qualitative Changes in Bath ViscosityCaF2 ↘↗, first down, then upAlF3 ↓LiF ↓

MgF2 ↑NaCl ↓NaF ↗↘, first up, then down

Al2O3 ↗, not linearlyTemp ↓

Metal Solubility

The main cause of the loss in current efficiency is the so-called back reaction, where dissolved aluminum in the bath reacts with dissolved CO2 producing alumina, CO and releasing heat according to the reaction:

Decreasing the amount of dissolved aluminum into the bath is one of the mean to increase the current efficiency. This can be done taking into account that the saturation concentration for the dissolved aluminum in the bath is affected by the electrolyte composition and its temperature.

The following table contains the qualitative effect of the bath additives on the metal solubility:

Increased Bath Component Qualitative Changes in Metal SolubilityCaF2 ↓AlF3 ↓LiF ↓

MgF2 ↓NaCl ↓NaF ↑

Al2O3 ↓Temp ↑

Surface Tension

The surface tension of the bath affects the dimensions of the CO2 bubbles. So, having a high surface tension implies to have bubbles with reduced dimensions and hence making more difficult the diffusion of the CO2 into the bath, with an increase in current efficiency.

The following table contains the qualitative effect of the bath additives on the bath surface tension:

Increased Bath Component Qualitative Changes in Bath Surface TensionCaF2 ↑AlF3 ↓LiF ↑

MgF2 ↑NaCl ↓NaF ↑

Al2O3 ↘, not linearlyTemp ↓

Vapor Pressure

The bath vapour pression is a measure of the bath volatility, hence affecting the amount of gas evaporated from the molten bath during electrolysis.

Again, the following table summarizes the quantitative effect of the bath additives to the vapour pressure:

Increased Bath Component Qualitative Changes in Bath Vapor PressureCaF2 ↓

AlF3 ↑LiF ↓

MgF2 ↓NaCl ?NaF ↓

Al2O3 ↓Temp ↑

Current efficiency

In an alumina reduction cell the aluminum is produced by the following reaction:

driven by the electrical energy given from the external of the cell.

In parallel with this reaction, another reaction occurs which reoxidizes part of the aluminum produced to alumina, according to the reaction:

This reaction is usually called “back reaction” and is the main factor driving a current efficiency less than 100%.

Let’s see now more in detail how the back reaction works and what can be done in order to minimize it, hence increasing the current efficiency.

The picture above describe in some detail the mechanism of the back reaction:

Some metal diffuses from the aluminum pad into the molten bath

At the same time some CO2 diffuses into the bath

Aluminum and CO2 diffused can than react according to the reaction (2)

To minimize the back reaction we need to work on the factors that reduce the mass transfer of aluminum into the molten bath. In particular:

Reduce the bath temperature, because this reduces the aluminum solubility

in it

Increase the bath acidity, because this reduces the aluminum solubility in it

Keep the interface of the metal pad as flat as possible, hence reducing the

cell noise

Optimizing the magnetic design of the pot in order to decrease the metal

velocity

Keep an adeguate ACD, in order to have more space between the zone

where the dissolved aluminum is present from the zone where the dissolved CO2 is present

The back reaction is the most important factor reducing current efficiency, but it is not the only ones. Some other factors reduce the current efficiency:

In presence of a vigorous agitation of the metal pad due to unbalances in the

anodic current distribution, metal waves can touch directly one or more anodes. In this case the

electrical current passes directly from the anode to the metal pad without producing electrolysis of

the alumina and hence reducing the current efficiency

During an anode effect the alumina reduction reactions are interrupted. The

electrical current flows without producing aluminum and hence reducing the current efficiency

The cryolite ledge

In an aluminum electrolysis cell the electrolyte is found in a molten state with a

working temperature only 10 ÷ 20 °C above the freezing point, while the molten aluminum of the metal pad is 300 °C above its freezing point.

Because the electrolyte is so close to its freezing point, we find a ledge of frozen electrolyte attached to the sides of the cell, being these ones the coldest area of a pot.

This frozen electrolyte ledge is made only of cryolite. In fact, looking at the phase diagram of the system NaF – AlF3, we see that when the electrolyte freezes the first phase which is separated is the cryolite.

The ledge plays an important role in the heat balance and bath chemistry control of a pot:

Heat balance: the thickness of the cryolite ledge changes according to the thermal

behavior of the cell. In fact, whenever the temperature of the electrolyte increases with an increase in the

temperature difference between the bath temperature and the liquidus temperature, some ledge melts.

The melting of ledge happens at constant temperature, thus “absorbing” heat and decreasing the

temperature excursions. Similarly, whenever the difference between the bath temperature and the

liquidus temperature decreases, some bath freezes increasing the ledge thickness and helping the cell to

reduce the temperature excursion

Because the ledge is made only of cryolite, the melting and freezing of it will change

the bath composition, primarily the ratio (or, in other terms, the AlF3 excess) because some liquid cryolite

will be added or subtracted from the molten bath.

The cryolite ledge protects also the carbon side of the cell from the chemical attack of the electrolyte. In fact, without this protecting layer, the molten electrolyte in direct contact with the carbon sides would destroy them very quickly, leading to bath tap-out and loss of the cell

Cell Thermal Balance

In an aluminum reduction cell the alumina is reduced to aluminum according to

the following reaction:

The energy requirement for this reaction is equal to the change in enthalpy between the products and the reactants. In our case:

Assuming a 100% current efficiency and being included also the heat necessary to increase the temperature of the reactants from the room temperature up to the operating temperature of the cell. More in general, considering a real case where current efficiency is less than 100%, we have:

Where CE is the current efficiency expressed as a fraction between 0 and 1. As an example, for a normal current efficiency of 93% we have ΔH=6.45 kWh/kg.

Let’s now calculate the energy requirements to produce aluminum expressed in terms of voltage. First, the specific energy requirement expressed in Joule is equal to:

While the production rate of aluminum per second, given a line current of I amperes and a current efficiency equal to CE (expressed as fraction of 1 and not in percentage), is given by:

Multiplying and rearranging the two equations we have the energy per second (hence, power) needed to produce aluminum:

This energy is supplied from the external as electrical energy:

or:

We have used the symbol VAl to say that this is the voltage required to produce aluminum.

The following chart depicts the equation above:

From the above calculations it turns out that the voltage required to produce aluminum is around 2V. But we know that a pot works with a voltage higher than this, typically 4.0 ÷ 4.5 volts. The difference between the voltage of a real pot and VAl is equal to the energy that is lost in the ambient as heat. This is also the heat that keeps melted the aluminum and the bath in the pot.

Let’s consider the various voltage components:

where:

Erev is the voltage to apply in reversible conditions for the basic reaction to occur

The terms indicated with η are the so called “overvoltages”:

ηcc: concentration overvoltage at cathode

ηaa: concentration overvoltage at anode

ηac: reaction overvoltage at anode (0.6 ÷ 0.9 V)

The terms indicated with the letter V denotes ohmic voltage drops. In detail:

VA is the voltage drop in the anodes

VB is the voltage drop in the electrolyte

VC is the voltage drop in the cathode

VX is the voltage drop in the buss bars external to the pot but which

contributes to the total pot voltage

The voltage components that are located entirely into the ACD space are:

The amount of voltage generated into the interelectrode space and used to produce aluminum is equal to VAl. So, the amount of voltage that generates heat that must be subsequently dissipated into the external environment is equal to:

The terms VA, VC and VX represents too heat that needs to be dissipated outside from the pot, but they are constant, do not change over a short time and are located outside the interelectrode space.

This heat leaves the pot basically following two main paths:

1. Through the anodes and the cathodes, following a vertical path

2. Through the cryolite ledge, following an horizontal path

Let’s analyze more in detail these two heat fluxes, starting from the one following a vertical path.

The heat exiting the cell upward has to go through the anodes and the crust covering them. Some fraction of this heat is released also through the anode stubs. The heat exiting the cell downward has to pass through the metal (with a thermal resistivity which is negligible compared to the resistivity of all the other materials present in a cell) then the cathodes, the bricks layers and the potshell. What is important to understand is that the thermal resistivity of these parts as well as the areas crossed by this heat do not change in the short period:

with:

RCathode: thermal resistivity of the cathodes

TM: metal temperature

TA: ambient temperature

TB: bath temperature

ACathode: cathode area

RAnode: thermal resistivity of the anodes

AAnode: anode area

Because the metal temperature is close to the bath temperature, we have:

And we can express the heat escaping from anodes and cathodes as:

Because, under normal operating conditions, the terms ACathode, RCathode, AAnode, RAnode do not change and the interval of operating cell temperatures (Tb) is not so big, we can say that:

For the heat exiting the pot from the cryolite ledge (horizontal path) we can write:

with:

α: liminar coefficient of heat exchange between electrolyte and ledge

TLiq: electrolyte liquidus temperature

ALedge: ledge area

RSide: heat resistance including cryolite ledge + cell side insulations + potshell +

ambient air

Because QV is approximately constant, all the variations in the heat generated by Joule effect in the interelectrode space must be absorbed by the component QO.

Changes in the horizontal heat QO are automatically realized changing the area and thickness

of the cryolite ledge.

Anode effect

The anode effect is a particular working state of the cell characterized by:

High voltage

Interruption of the aluminum production

Production of green house gases CF4 and C2F6

and usually is triggered by a decrease in the alumina concentration in the bath below 2%.

In the ’60 Piontelli found out that the anode effect starts whenever in one of the anodes of a cell the current density exceeds a critical current density whom value depends upon the bath temperature and the alumina concentration.

During an anode effect the normal reactions that produce aluminum are interrupted and other electrochemical reactions take place with the formations of gases CO, CF4 and C2F6. These gases forms bubbles that adhere to the anode bottom, creating an electrical insulating layer which is the reason why the voltage of the cell increases up to 10 – 50 volts and even

more. The maximum voltage reached during the anode effects depends also by the bath level. With high bath levels the maximum voltage reached is lower because some fraction of the line current passes through the sides of the anodes, free from the adhering bubbles.

During an anode effect it is possible to observe on the anode surface small sparks that make visible how the current passes from the anodes to the bath through the gas bubble adhering to the anodes.

In order to turn off an anode effect:

The alumina concentration must be restored to normal values as quick as possible

The layer of gas adhering to the anode bottom must be removed

In a modern pot with automated control system and point feeders, the computer detects an anode effect when the voltage raises above 8 V. After the detection the control systems starts an overfeeding of the pot, which means to activate at a very fast rate for some time the pot feeders, and, if enabled, try to remove the layer of adhering gas through particular up and down movements of the anodes.

To remove the layer of adhering gas it can be used also green poles which are inserted into the molten aluminum pad. The green poles provoke a violent stirring of the bath and metal and this stirring remove the adhering layer of bubbles.

In the last decades the aluminum industry has greatly reduced the anode effect frequency because of the CF4 and C2F6 green house gases generation, which have a global warming potential of 6500 and 9200 respectively. The ratio of the mass of produced gas CF4/C2F6

during an anode effect is equal to 10.

Influence of Magnetic Fields

When an electrical charge moves in a magnetic field, the magnetic field itself exert

a force on the moving charge.

In the alumina reduction cells we have strong magnetic fields, generated by the high intensity electrical currents used for the electrolysis (now going beyond 350000 Amps and in some test cell 500000 Amps), and electrical charges (electrons in the metal pad and ions inside the electrolyte) moving in these magnetic fields.

The force exerted by the magnetic field on the electrical charges moving is calculated by a vectorial product:

Where B is the magnetic field while j is the current density. The direction of the force F will be perpendicular to the magnetic field B and the current density j. The direction of F follows the right-hand rule and its unit measure is N/m3.

As a result, these forces induce in the electrolyte and in the molten metal complex movements, in terms of vortexes and bath/metal interface waves. All these movements have a detrimental effect on the current efficiency and the specific energy consumption therefore it is of paramount importance to design pots with good magnetic compensation in order to decrease to a minimum all the bath and metal movements.

Let’s consider one cell in a side by side configuration. To help studying the forces that develop inside the electrolyte and metal, it is useful to refer to a right hand coordinate system with the origin O at the cathode center, the plane defined by the origin O and the axes X and Y parallel to the cathode, the X axis aligned as the line current, and the Z axis going out of the cell through the molten metal and then the electrolyte:

Taking this coordinate system as reference, we can split the magnetic field B and the current density j in their components Bx, By and Bz and jx, jy and jz and list in the following table the

various components of the forces generated by each B,j components couple:

Magnetic Field Component

Current Density ComponentHorizontal

Transverse Jx

Horizontal Longitudinal Jy

Vertical Jz

Horizontal Transverse Bx

0Vertical Force Fz

(-)Longitudinal Force Fy (+)

Horizontal Longitudinal By

Vertical Force Fz

(+)0

Transverse Force Fx (-)

Vertical BzLongitudinal Force Fy (-)

Transverse Force Fx (+)

0

Inside the electrolyte and/or the molten metal, it is possible to have vortexes only if the following integral:

(where w are the forces arising from the gravity) calculated in a closed path is greater than zero.

To avoid any movement in the molten medias the forces need to be balanced point by point per cell quarter as in the drawing below:

To achieve the force equilibrium as outlined in the picture above, for every set of four points K1, K2, K3 and K4 (symmetrical respect to the XZ and YZ planes) the magnetic field components must be antisymmetric, or, in other terms:

Bx(K1) = -Bx(K2) = Bx(K3) = -Bx(K4) (1)

By(K1) = -By(K2) = By(K3) = -By(K4) (2)

Bz(K1) = -Bz(K2) = Bz(K3) = -Bz(K4) (3)

These equations imply that along the X and Y axis the magnetic field components are equal to 0.

For the usual arrangement of a side by side cell the current conductor and the ferromagnetic masses are arranged in a way that the Bx and By components are antisymmetric, hence automatically fulfilling the conditions (1) and (2).

This means that the vortexes generation depends mainly on the value of the Z component of the magnetic field. To avoid any vortexes the current conductors and ferromagnetic masses should be arranged in a way that the Bz component is antisymmetric point by point as expressed by the (3) or at least antisymmetric in average in the four quadrants.

If any of the magnetic field components do not satisfy the (1) and/or (2) and/or (3) then the electrolyte and the metal start to move.

Because the Bz component acts together with the horizontal components of the current density, to limit the forces that set movement into the molten medias it is important to reduce as much as possible the jx and jy current density components.

Inside the electrolyte jx and jy are close to zero, and the current is mainly vertical, while in the molten metal the following causes:

Unbalanced currents between the different anodes

Muck and/or crust in the bottom

Cryolite ledge too much thin or terminating below the anodes

lead to horizontal current density components with subsequent set up of an oscillating wave in the metal pad.