NO168432B - DRAINAGE CATODECEL FOR ELECTROLYTIC REDUCTION OF ALUMINUM OXYDE TO ALUMINUM - Google Patents

Description

The present invention relates to a new drainage cathode cell for the electrolytic reduction of aluminum oxide to aluminum in a cryolite-based bath containing aluminum oxide. More specifically, the invention relates to an aluminum-producing electrolysis cell where the bath electrolyte is based on sodium cryolite, where the problems caused by reduced anode-to-cathode distance that occur in the previous drainage cathode cells are solved by inducing a special type of bath current in the ACD gap where the supply of aluminum oxide is facilitated, in the same way as the removal of gaseous products, and it is made easier to tap off the produced metal. Operation of the experimental electrolysis cell has shown that it is possible to provide an abundant supply of dissolved aluminum oxide to the electrolysis zone, even with a very small distance between anode and cathode.

A common type of electrolytic cell for the production of aluminum is the classic Hall-Heroult design, which uses carbon anodes and an essentially flat carbon-lined bottom that functions as part of the cathode system. An electrolyte is used in this production of aluminum by electrolytic reduction of aluminum oxide, the electrolyte mainly consists of molten cryolite with dissolved aluminum oxide,

and may contain other materials such as e.g. calcium fluoride, aluminum fluoride, and other metal fluoride salts. Molten aluminum resulting from the reduction of alumina is usually allowed to accumulate at the bottom of the vessel constituting the electrolytic cell, as a pad or pool of molten metal over the carbon-lined bottom, thereby forming a liquid metal cathode. Carbon anodes that extend into the container and come into contact with the molten electrolyte are aligned relative to the liquid metal cathode. Current collector bars, such as e.g. steel, is most often cast in

in the carbon-lined cell bottom, and completes the connection to the cathode system.

Although designs and sizes of Hall-Heroult electrolytic cells vary, all have a relatively low energy efficiency, ranging from 35 to 45% depending on cell geometry and operating conditions. While the theoretical power requirement for the production of 1 kg of aluminum is approx. 6.3 kWh, the power consumption in practice is in the range from 13.2-18.7 kWh/kg, with an industrial average of approx. 16.5 kWh/kg, with an industrial from theoretical energy consumption is due to voltage loss in the electrolyte between anode and cathode. Accordingly, the possibility of reducing the anode-cathode distance (ACD) has been studied.

Because the molten aluminum pad that serves as the cell cathode can be irregular and variable in thickness due to electromagnetic effects and the circulation in the bath, past practice has prescribed that the ACD should be kept at a safe value of 3.5 to 6 cm, so that one is' ensured a relatively high current efficiency, and avoids a direct short circuit between the anode and the metal pad. Such gap distances give a voltage drop of 1.4-2.7 V, which is in addition to the energy required for the electrochemical reaction itself (2.1 V, based on enthalpy and free energy calculations). Consequently, great effort has been directed towards developing a more stable aluminum pad, so that the ACD can be reduced to less than 3.5 cm with associated savings in energy.

Refractory, hard materials (RHM), such as e.g. titanium diboride, has been studied for some time for use as cathode surfaces in the form of bricks, but until recently adherent RHM bricks or surface coatings have not been available. It is known that titanium diboride is conductive, it also has the unique property that it is wetted by liquid aluminium, thus enabling the formation of very thin aluminum films. The use of a very thin aluminum film flowing down an inclined cathode covered with a RHM surface, to replace the unstable molten aluminum pad of the prior art,

is proposed as a means of reducing ACD, thereby improving efficiency, and reducing voltage loss. Previous attempts have not succeeded, however, because the available RHM surfaces have not been adequate, and because the problem of providing a sufficient supply of dissolved alumina to the

limited ACD (as small as 1.5 cm). The problems with insufficient supply of alumina occur at minimal ACD, and include large and persistent anode effects. Overfeeding alumina to avoid these problems results in sludge deposition, which can clog the cell and prevent its operation.

An alternative method to reduce energy consumption has been to melt aluminum from aluminum chloride instead of aluminum oxide. This process for producing aluminum requires 30-40% less electrical power than conventional electrolysis. In this method, the usual Bayer process is used to convert bauxite to aluminum oxide, which is then converted to aluminum chloride in a chemical plant, and then melted in an electrolytic cell. In the cell, aluminum chloride is broken down into aluminum, which is drained off, and chlorine, which can be recycled back to the chemical plant for the production of more aluminum chloride. Such techniques use a flow-through reactor which

have permanent anodes. The aluminum-producing cells in this method are, however, incompatible with Hall-Heroult-type cells, and cannot be retrofitted to an existing aluminum plant. The chloride process therefore requires large investments in a completely new installation.

The present invention enables significant energy savings to a greater extent in a retrofitted facility, and removes the need for completely new installations. • It is against this background that the present invention was developed.

It is an aim of the present invention to provide an improved aluminum electrolysis cell, where the bath electrolyte is based on sodium cryolite, which has improved electrical efficiency. It is thus an aim of the present invention to provide a basic cell for the production of aluminum metal from aluminum oxide which can be used in a single-cell system, or a multi-cell system, which is economically advantageous both with regard to construction and operation.

It is a purpose of the present invention to provide a cell design that provides a narrower ACD than the previous designs, and thus improved current efficiency and voltage drop, which can be retrofitted to existing Hall cell installations. Consequently, there is a purpose

by the present invention to provide an electrolysis cell with an ACD of less than 3 cm, where a sufficient supply of aluminum oxide to the anode-cathode gap is ensured, without requiring an oversupply. Further purposes of the invention are to provide a design of an aluminum reduction cell which includes a flow-through configuration and a controlled atmosphere.

These, as well as other purposes, which will appear more clearly on the basis of what follows, are achieved by providing a cell where the individual process steps are separated from each other to the greatest extent possible. An inclined, fixed cathode is used to form the anode, which can be either of the Søderberg type or burned, which, together with the appropriate choice of other parameters defined here, induces such a bath current through the anode-cathode gap that the proper supply of alumina to the anode is provided, without adverse effect on the tapping of the aluminum product in a collection well. In one configuration, the gas pumping action below the anode causes the bath to flow through the cell, into a supply chamber for the cell, out of the supply chamber, and back to the opposite side of the cell. Several individual cells of this type can be combined in numerous configurations, so that the system can be structured for the physical and economic restrictions of any facility.

Accordingly, the present invention provides a drain cathode cell for the electrolytic reduction of alumina to aluminum in a cryolite-based bath containing alumina, characterized in that it includes:

a shell having an inner surface lined with refractories and carbon materials so as to define a cathode cavity; a cathode having at least one upper surface containing a hard aluminum wettable refractory material; at least two anodes extending into the cathode cavity, both of which have a lower surface spaced from said upper surface of the cathode at an anode-to-cathode distance of 1 cm to 5 cm, defining an anode-cathode distance ( ACD) and a flow path between anode and cathode where the upper cathode surface and said lower anode surface are inclined from the horizontal plane by from 2° to 15°, so that a lower end and an upper end of the flow path between said anode and said cathode are provided; a device in the cavity to provide during operation a minimum bath flow rate Q(cell) around at least one substantially horizontal bath circulation circuit to ensure supply of alumina for the electrolysis reaction wherein (where Awt% A1203 is the difference in wt% AI2O3 in the bath at the entrance and exit of the ACD gap) by providing at least one return flow path to complete the loop, the loop including: (i) said flow path between the anode and the cathode; (ii) at least one lower channel in fluid contact with the lower end of said flow path between the anode and the cathode; (iii) at least one upper channel in fluid contact with said upper end of the flow path between anode and cathode; (iv) at least one return flow channel in fluid contact with said upper channel and said lower channel; the return flow channel, placed between neighboring anodes, has dimensions h, w and L, where h is the bath depth in cm at any given point, w is the width of the return channel in cm, and L is the length of the return channel in cm, and where h , w and L are determined by the relation where Rf is the flow resistance in the return flow channel, and in which the maximum value for Rf (cell channel) is determined by the relations

and wherein (X) ranges from 0.2 x 10"<3>/cm<4> to 230 x 10-<3>/cm<4> and where (X) is a geometric resistance factor.

The present invention thus relates to an electrolytic cell in which the cathode is so arranged and has such a slope that the anode gases induce an abundant and sufficient flow of enriched materials between the anode and the cathode, without preventing the draining of molten aluminum into a collecting well, or destroying the surface of that aluminum which is drained off so that a lower power efficiency is caused. Consequently, the bath is caused to flow from an enriched zone to a zone where the bath is somewhat depleted in aluminum oxide content. The bath can be equipped with an automatic supply of aluminum oxide in a controlled quantity, to prevent soiling or anode effects. Figure 1 shows an electrolytic cell having an inverted V-shaped cathode, or double-sloping cathode configuration, viewed from the end. Figure 2 shows a cell concept according to the present invention,

with single-sloped cathode, viewed from the end.

Figure 3 shows a flow-through aluminum reduction cell with a single sloped cathode and outer filling viewed from the end. Figure 4 shows a V-shaped drain cathode cell seen from the side

the end.

Figure 5 shows a draft of a single sloped drain cathode cell, which uses a system for continuous aluminum draining and separate filling and mixing zone. Figure 6 shows a draft of a multi-cell system, which uses several single drain cathode cells in tandem. Figure 7 shows a draft of a single slope drain cathode cell that has partial barriers that limit the bath flow rate. Figure 8 shows a draft of a cell that has two opposite sides

inclined cathodes.

Figure 9(a) shows the rise in a hydraulic analogue model,

seen from the side, used to simulate gas evolution in a sloping cathode cell.

Figure 9(b) shows normal bath flow in ACD.

Figure 9(c) shows an area with reversed bath flow in ACD,

near the aluminum metal drain.

Figure 9(d) shows an area with turbulent and interrupted bath flow,

where there is no net flow of the bath material through the ACD.

Figure 10 shows a graphical presentation of bath current as a function of cathode slope in the cell simulation model under conditions with high flow resistance. Figure 11 shows a graphical presentation of bath current as a function of cathode slope in the cell simulation model under conditions with medium flow resistance. Figure 12 shows a graphical presentation of bath current as a function of cathode slope in the cell simulation model under conditions of low flow resistance. Figure 13 is an example of a pump efficiency-flow-resistance diagram for a drained cathode cell.

Figure 14 graphically shows flow resistance in the return channel

as a function of cathode slope.

Figure 15 constitutes a cell design parameter diagram.

Figure 16 is a cross-section of an experimental electrolysis cell

with sloping cathode.

Figure 17 is a graphical representation of the operational stability

obtained in three experimental cells.

Figure 18 shows the relationship between the anode-cathode polarization and the cathode slope at two current densities.

The present invention consists in the use of an inclined, drainage cathode, which has a surface of titanium diboride or a similar hard refractory material, where the slope or deviation from the horizontal plane is chosen so that it contributes to a controlled movement of the bath, where the bath movement is hydrodynamically dependent on specific identified critical parameters. Bath movement is controlled

so that it causes sufficient alumina-rich bath to flow into the space between the electrodes in the space just below the anode surface, so that anode effects are avoided, while at the same time the draining of aluminum is not affected. It is thus possible to overcome the problems with alumina supply, which e.g. fouling due to excessive addition to the bath, persistent anode effects due to insufficient addition, and the formation of protruding edges, affecting bath circulation, which occur at very low values of the anode-cathode distance (ACD), made possible by the development of cathodes with surfaces of hard refractory material which is used without the conventional metal pad.

The success of a low energy aluminum reduction cell with a drain cathode depends on two critical components: 1) a durable RHM cathode material, and 2) the ability to control and optimize the bath current through the narrow anode-cathode gap (ACD). Ongoing trials in production cells indicate that a material has been developed that satisfies the first requirement.

Studies where a water model has been used to simulate

the hydrodynamic conditions in ACD in an inclined cathode cell, have now provided criteria for cell design necessary for the second requirement to be met. Results from

the water model has been verified in a laboratory-sized electrolysis cell. A number of cell designs, such as in the experiments with the Kaiser-DOE inclined TiB2~cathode reported under contract DE-AC03-76CS40215, and used in other published reports and patents, fall outside the present design requirements, and fail to address crucial hydrodynamic issues, the scope of which is now being seen for the first time . These design criteria include the slope of the cathode surface, the ACD spacing, the flow resistance of the bath current on the return path, and the anode current density. Only the first two of these criteria have been considered in previous designs, reports and patents, while the last two criteria, and their importance, are established for the first time here.

It is rooted in aluminum electrolysis that the anode surface

during the electrolysis is set so that it is approximately parallel to the cathode. A sloping cathode causes the anode to have a corresponding slope. As is well known, there is abundant gas evolution on the anode surface due to the electrolysis process.

It is also well known for aluminum reduction cells of conventional design, with horizontal anode and cathode surfaces, that gas developed on the anode surface goes obliquely (laterally)

out below the anode surface from the center of the cell to the nearest vertical vent, the vents generally including the vertical side surfaces of the anode. A sloping anode can either facilitate or hinder this gas movement, depending on the direction and magnitude of the slope. A suitable cathode, and therefore anode, slope can consequently cause the anode gas to move in a desired direction. Furthermore, these anode gases will drive the bath near the anode surface in the same desired direction, by a well-known effect which is also utilized in the design of gas aerators and gas jet pumps, as well as bubble columns. This operation can be supplemented and increased (if necessary) by other pumping action, e.g.

using a pump of suitable design.

It is well known in hydrodynamics that the flow in

a fluid system arises from a balance between the drift of the fluid and the resistance to flow within the parts of the system, and that, depending on the configuration, the velocity within local areas of the flow may be in the same direction as, in some cases oppositely directed, the direction of the fluid drift. It is a principle of the present invention that the slope, or slopes, are arranged so that a balance is achieved between the forces that drive the buoyancy bubbles from the slope and the forces that drive the bubbles laterally under horizontal anodes on the one hand, and the flow resistance on the other side, such that a net movement occurs in the bath which provides the aluminum oxide supply required. The local velocity in the bath near the anode surface is in the desired direction, i.e. in the direction in which the drifting gas bubbles move. As before, the configuration is positioned to provide bath velocities near the cathode surface in the same direction as near the anode, not so high as to interfere with the stripping of aluminum onto the cathode,

where the aluminum draining takes place in the opposite direction to the flow of gas bubbles that drive the bath. With the alumina-rich bath source located on a particular side of the cell, a bath current is obtained into the space immediately below the anode surface when this side of the anode is low and the surface slopes upward from this side. The slope must be sufficiently large to overcome, and reverse, the flow of gas that would otherwise go towards this anode edge from the inner parts of the anode. The exact configuration and choice of slopes can be varied by the location of the source of the alumina-rich bath. This can be on one side, on both sides, or in the center of the cell, depending on the size of the cell and the type of anode used (burnt or Søderberg).

One can imagine alternative embodiments of the invention where the cathode (and therefore anode) slope is uniform across the width of the cell, or where the slope is variable, in varying forms. These can e.g. include various,

but constant slopes in the same direction in the two halves of a transverse section through the cell, and also equal but

opposite slopes in these two halves (eg a double-sloped, inverted V configuration). The concept in this invention can be used for cells that have both burnt and/or Søderberg anodes. While e.g. the inverted V configuration may be most suitable for cells where the width is covered by two burnt anode blocks, such a configuration can also be used with a Søderberg anode in the form of a vertical bolt with a vent hole in the centre. In the case of an inverted V configuration, the supply to the cell can take place from both sides, while a V configuration is most suitable for supply in the centre. Correspondingly, a design with an inclination in only one direction (constant or variable) can preferably be used, but the invention is not limited to a simple, monolithic, unventilated Søderberg anode where the supply takes place from one side.

The choice of cathode slope for a particular application must be made compatible with other important parameters so that the desired hydrodynamic characteristics are achieved. These parameters are the ACD gap, the anode current density, (current and anode surface dimensions), and the return resistance of the bath (ie, the bath return channel or passage length, depth, width, and wall material). The controlled bath current ensures a sufficient supply of aluminum oxide-rich electrolyte to the anode surface so that large anode effects are avoided. Further advantages are that excessive gas accumulation in the ACD is avoided, reduction of the anode and cathode overvoltage, and that the flow of the aluminum product is disturbed to a minimal degree.

Figure 1 shows a Søderberg reduction cell with vertical bolt and side intake, seen from the end, which uses a cathode made in the form of an inverted V with a central gas discharge,

and busbar. The cathode surface, 6, slopes upwards from the cell sides towards the center of the anode, 1, where the anode gas is released through the gas hole, 2, to a collection system (not shown). The busbars, 4, are shown approximately parallel to the cathode surface, 6, and are retracted to allow aluminum to pass to the collection well 5.

The angle the cathode surface 6 forms with the horizontal plane

should be large enough to induce the desired bath current, but not so large as to cause significant deformation of the bubbles, or bath turbulence in the anode-cathode space, 7, or other limitations described below. It is e.g. observed that the angle of inclination has a pronounced,

and previously undescribed effect on bubble configuration and motion. When the anode surface is inclined in relation to the horizontal plane, the bubbles show a greater tendency to an elongated oval shape, where the long axis is perpendicular to the direction of movement of the bubbles upwards: the anode slope. The leading edge of the bubbles forms a crest, or increased thickness,

when the relative velocity between the bubbles and the liquid increases. The thickened leading edge can provide less driving force

for the bubble propulsion over a certain range of helninqs angles, due to increased resistance due to displacements and friction at the interface. It has been observed that large bubbles on the anode surface are exposed to significant distortion and resistance at an anode slope of approx. 15° from the horizontal plane. Suitable tilt angles have been found to be in the range from 2 to 15°, although somewhat larger or smaller angles may be usable for given cell conditions. A preferred inclination is from 5° to 10° from the horizontal plane, a more preferred inclination is from C to 8°. The most preferred cathode slope is found to be approx. 8°.

The cathode surface, 6, can be covered with an electrically conductive material that is wetted by aluminium, such as e.g. TiB,,, to facilitate the formation of a thin film of aluminum (or aluminum alloy) on the cathode surface. While surfaces containing titanium diboride are preferred, the use of other hard, refractory materials (RHM) wetted by aluminum has been considered, such as titanium carbide, zirconium carbide, zirconium diboride or mixtures thereof.

Aluminum flows from the sloping surface of the cathode

into a collecting drain located on the side of the cell,

5, thereby minimizing the probability of a back-reaction that would reduce the current efficiency, and the tapping process

the animals become faster, and independent of other cell operations. This thin film of flowing aluminum is insensitive to induced magnetic fields, a major improvement over the previous conventional pads of molten aluminum. Furthermore, the controlled bath movement,

as described above, do not inhibit the draining.

The gas bubbles, 8, which move along the lower sloping edge of the anode, are aerated as needed, so that the bubbles do not become excessively large, i.e. so that they do not cover a significant part of the anode surface, or have a very large extent in the space between anode and cathode. However, overaeration should be avoided, so that the driving force for the bath flow is not significantly reduced in the narrow area between anode and cathode. Venting can be done through a slot or slots or through ventilation pipes, ducts, holes, etc.,

in the anode. Venting holes can be made by inserting pipes through the electrode mass or the plastic zone of a Søderberg anode, or baked into a burnt carbon anode. It should be noted that a non-uniform distribution of the air holes in the anode will give a new distribution of the bath current in the cell,

to compensate for the non-uniformity of the bath flow.

Aluminum oxide can be added to the bath through funnel 19,

to replenish the depleted bath. Other components of the bath can of course be added at the same time. The arrow, 10, indicates the flow of alumina-rich bath through the narrow area between anode and cathode, 7, driven by the gas bubbles 8, while arrow 13 indicates the flow of aluminum metal,

as a thin film, to the drain well 5.

Figure 2 shows an aluminum reduction cell using a single sloping cathode, viewed from the end. The single-sloped cathode surface, 6, makes it possible to use a straight busbar, 4, which makes it possible to maintain a constant distance between the upper cathode surface and the busbar. Although there are certain advantages in terms of voltage reduction if the busbars extend from both sides of the cell, 22, this is not necessary if the deeper drain well for aluminum, 5, is more advantageous. The configuration of this cell makes it possible to use

more or less conventional construction and installation methods for placing the busbar. The slope of the cathode surface, 6, must be sufficient to effectively move the gas bubbles up along the anode surface, and to allow aluminum to flow down along the cathode surface. As shown, excess anode gas bubbles, 8, which could otherwise cause great distortion of the metal film flowing down the cathode fiat, can be vented through the vent hole in the anode, 2, to a gas collection system 3. By using a mechanical cover over the bath, 18, complete control over the bathroom atmosphere is ensured, and evaporation to the atmosphere is reduced. Aluminum oxide is added to the bath in the space between the electrodes, 21, using the funnel 19. An automatic control, 20, can be used to achieve exactly the right top-up. The addition of aluminum oxide, and other bath ingredients, in this way has several advantages. After it leaves the electrolysis zone, the bath is firstly partially depleted of dissolved alumina, thus providing ideal conditions for rapid dissolution of alumina. If any alumina settles to the bottom of the side reservoir, 24, the fact that there is no metal pad or well to cover the alumina will make it easier to dissolve it in the flowing bath. Furthermore, the return path for the bath provides additional time for dissolution and acts as a trap for undissolved alumina to ensure that only an uncontaminated bath circulates within the ACD gap. While the addition of aluminum oxide on the side of the cell where the metal drain well is located, 5, could result in the formation of contaminants under the metal in the well, thereby causing flooding and the possibility of direct short-circuiting at the metal between anode and cathode, this is avoided with the present solution.

Figure 3 shows a flow-through aluminum reduction cell with a single-sloped cathode viewed from the end. As shown, the cathode angle is sufficient to effectively move the anode gas bubbles, 8, up along the anode surface, 1, and to allow the aluminum to flow down the cathode surface, 6, to the aluminum drain well,

5, as shown by arrow 13. The vents, 2, which may be required, are shown in the anode, and may be a slot or a pattern of vents. The anode itself can be of the Søderberg or burnt carbon type. The drain well for aluminum,

5, preferably leads to a continuous tapping system,

which causes the least possible disturbances in the cell, such as e.g. variations in pad thickness. The air holes, 2, and empty cavities in the bath, can advantageously be connected with a gas collection system, 3. In the flow-through cell, the electrolysis bath is replenished as shown from the outside, and flows through the bath inlet 9, the path shown by the arrows 10, through anode-cathode -compartment 7, to the bath outlet 11, at this point the electrolyte has been partially depleted of aluminum oxide by the electrolysis.

Figure 4 shows a double-sloping drainage cathode, or V-shaped cathode. In this configuration, replenishment is provided by the point feeders 27, which are located at the highest end of the cathode slope, although replenishment can also take place between the burned anodes, 1. The flow of aluminum metal, 13, is down the cathode surface 6 to the drain well 5, which can advantageously be connected to a continuous draining system. The anode gas bubbles 8 cause a flow of bath material 10 along the lower surface of the anodes.

Figure 5 shows a sketch of a single cell, which includes filling and mixing zones. In figure 5, the anode 1 is equipped with a gap for gas venting, 2. The cathode surface,

6, is shown with the aluminum drain well, 5, feeding a continuous draining system, 14. The direction of the net flow of aluminum metal is shown by arrow 13, while the direction of the net flow of the cryolite in the bath is shown by arrows 10. After passage between the anode 1 and the cathode surface,

6, the partially depleted bath flows from the outlet 11 to a feed and mixing area or zone, 16, where alumina is replenished, 12. After replenishment of alumina, the cryolite bath is recirculated in the electrolytic cell in the direction shown by the arrows 10. The drawing shows the use of an optional magnet pump for bath circulation, 17, located near the inlet 9, although such a pump is not considered

as necessary under most conditions. A barrier,

15, can be placed in the return channel, 28, to prevent backflow and interruption of circulation in designs that include a supply and mixing zone (16).

Figure 6 shows a sketch of a multi-cell system, which consists of four individual electrolysis cells operated in tandem. The system shown includes a continuous aluminum drain system, 14, which provides a constant aluminum level, and consequently a more stable heat balance and a more stable bath level. Conventional discontinuous metal tapping could have been used, but precautions had to be taken to compensate for the resulting changes in the bath level. This would make it necessary to include a swing chamber for the bath, a lower level in the aluminum supply/mixing tank, and a device to regulate the heat balance. As shown in Figure 5, aluminum oxide and other bath ingredients, 12, are added to a supply and mixing zone,

16, and is circulated through the anode-cathode gap in the direction shown by arrows 10. Use of a magnetic circulation pump, 17, is optional. The formed aluminum flows into the aluminum drain wells, 5, and passes in the direction shown by the arrows 13 to the continuous metal tapping system 14. Backflow of the depleted bath is prevented by the barriers 15.

A system for supplying and dissolving aluminum oxide

can be easily incorporated into either a single-cell system or a more economical multi-cell system. A feature of the present system which enables it to be operated under a controlled atmosphere is a sealed cell top, thereby reducing material problems in the reactor, problems in controlling the drawdown, and heat losses, although a sealed top is not required. If this design is equipped with anode surfaces that do not corrode, the result is a complete, unattended, multi-cell unit in a single vessel.

Figure 7 shows an aluminum reduction cell, which uses a single sloping cathode. The figure shows a continuous draining system for aluminum 14, which removes aluminum from the cell via the drain well 5. This provides a constant aluminum level in the cell and also results in a more stable heat balance and a more stable bath level in the cell.

An adjustable partial barrier, 25, is used if necessary to limit the circulation rate of the bath in the cell in the return channel 28. The desired limitation can be achieved by means of barrier plates, consisting of a non-corrosive insulator, such as e.g. silicon nitride. Necessary means for filling are not shown, but may be of any suitable type as discussed earlier.

As previously indicated, a double-sloping cathode cell can alternatively be used within the scope of the present invention, as e.g. has an inverted V configuration,

as shown in Figure 1, or a V-shape, as shown in Figure 4,

where the metallic aluminum would flow down to the metal drain well. However, the single-sloped cathode, which slopes upwards towards the side of the cell where alumina is supplied, as shown in figure 2, can be considered a preferred design, since this design makes cleaning and other maintenance of the vessel in the cell much easier, while at the same time simplifying the cell construction itself .

Figure 8 shows a plan section through an aluminum reduction cell that uses double single sloping cathodes, this illustrates another alternative embodiment of the present invention. The flow of electrolyte is illustrated by the arrows 10, which indicate a circulation between the anode 1 and the cathode surface, 6, through a zone where replenishment, 12, takes place and through the ACD of neighboring anode-cathode pairs, which are inclined in the opposite direction. The center line of the cell, 26, can suitably be the boundary line for the two oppositely inclined cathodes. To ensure even and continuous circulation,

the barriers 15 and the partial barriers 25 are present to respectively prevent backflow of the bath,

and to control the circulation rate of bath current. All other elements in the drawing are as discussed earlier. It should be noted that this configuration represents an alternative to the previously proposed V shape

and inverted V-shape, and can be suitably used with burnt anodes rather than Søderberg anodes. Another design, which is particularly suitable for burnt anodes, has the inclination angles of neighboring cathodes (and anode surfaces) in the same direction, but the angles are of different magnitudes.

It should be noted that although the figures do not show presence

of side frames and/or crusts of frozen bath material,

then the present invention can also be applied to cells that include such.

The pumping action of the gas bubbles, directed upwards under a

gently sloping surface, where the liquid is in a thin layer between the upper surface and a parallel lower one, is first demonstrated in an analogous hydraulic experiment. The experiment was carried out in water at room temperature, whose kinematic viscosity is approximately equal to that of molten cryolite.

With a surface slope of 2.5°, and a distance between upper

and lower surface of 2.2 cm, bubbles were generated on

a porous upper surface using pressurized air,

with a gas evolution rate matched to that of a typical high-current aluminum electrolysis cell, this gave net or average liquid velocities in the space between the simulated anode and the simulated cathode in the range of 5-10 cm per Sec. It was calculated that these rates were more than sufficient to deliver alumina at the rate required for good operation of the cell at normal alumina concentrations in the cryolite. While the surface slope in this experiment did not have to overcome the oppositely directed bubble forces, which e.g. due to magnetic effects, which would be present in an aluminum electrolysis cell with a horizontal anode, the principle of gas-driven bath circulation was nevertheless clearly demonstrated.

It was found that the resulting bath circulation was controlled by the balance between the pumping effect of the bubbles in the ACD gap and the back pressure, or flow resistance, to bath circulation through the return channels. Additional results cg data were obtained in an aluminum electrolysis cell and a 1:20 scale hydraulic analog model.

The hydraulic analog model was constructed to simulate a gas-evolving anode surface (approx. 1.11 m2 ) placed over a sloping fixed cathode surface. This embodiment simulated an aluminum reduction cell with a typical "drain cathode" where the working anode is positioned over a sloping drain Til^ cathode surface. The studies of the water model were carried out with water at room temperature. The flow patterns and velocities observed in the water model were similar to those that would be expected in a full-scale cell, since the observed flows had Reynolds numbers in the turbulent region

(<>> 5000). In this region, the flow is mainly controlled by the physical dimensions of the flow channels and not by the fluid properties (eg viscosity). A compressed air chamber with a bottom consisting of porous Alundum plates (pore diameter about 20 (im) was used to simulate a working anode evolving gas (e.g., CO^). A gas velocity of 0.1778 cm/sec. through the porous anode plate was used to simulate an anode current density of 0.68 A/cm<2> (the gas velocity was corrected for differences in temperature and hydrostatic pressure). Simulated currents up to about 1.4 A/cm<2> were tested in this model.

The execution of the model, viewed from the side in Figure 9(a), simulated half of the cell shown in Figure 1. Viewed from above, the model simulated the cell shown in Figure 7. The end wall at the "upper" end of the cathode corresponded to a vertical plane that went through the center gap or the air hole, 2. The figure shows the connection between such parameters as e.g. ACD, BFL, <t>, h, hg, width of the lower channel, and width of

the upper channel (defined below) and the anode (1), the cathode surface (6) and the bath (29).

The following definitions are used to describe the parameters

in the water model and in full-size aluminum reduction cells, relative to the desired bubble flow upwards under the sloping anode surface: BFL - the dimensions of the anode surface in the direction of the bubble flow

(BLF is equal to 122 cm in most trials)

BFW - the dimension of the anode surface in a direction perpendicular to the direction of the bubble current (BFW is equal to 61 cm in most model tests)

ACD - the vertical distance between the inclined anode-

and cathode surfaces, corresponding to 7 in Figure 2 (ACD

was varied from 1 to 5 cm in the model tests) Cathode slope - the angle between the inclined cathode surface and the horizontal plane (the cathode slope was varied

from 0 to 15° in the model tests)

h - the vertical immersion depth of the anode from the liquid surface to the lower anode surface (h varies along

the BFL dimension of the anode depending on the cathode slope) hg - the minimum vertical immersion depth of the anode,

i.e. the anode immersion depth at the highest end of the cathode (h^ is equal to 10 cm in most model experiments)

The desired bath current typically passes through four different types of channels or passages, these are: 1) the ACD gap between anode and cathode, where the introduction and flow of anode gas generates the power needed to maintain the desired bath circulation

in the cell.

2) The bath current leaves the ACD gap and flows into the upper channel, this corresponds to reservoir 24 in figure 2, where the current is directed either towards one or towards both sides of the anode. Most of the anode gas is driven out of the bath in this channel, so that a turbulence effect is provided in the bath, which is ideal for dissolving the added aluminum oxide. This channel may be along the side of the cell which

shown in figures 2 and 7, or in the center of the cell as shown in figure 1. The dimensions of the upper channel are defined as: depth = depth of the bath in the channel (hQ + the vertical component of ACD in the water model, as at

low angles are approximately equal to ACD)

width = the horizontal dimension in the same direction

as the flow of bubbles in the ACD gap

length = the horizontal dimension above the BFW of the anode

plus the width of the return channel

3) The return channel or channels transport the bath from the upper channel to a corresponding lower channel located along the anode edge where the bath enters the ACD gap. The example of return ducts is marked with no. 28 in Figures 5-7. In its simplest form, and in the model used here, the return channel is as shown in Figure 7, where the depth and width of the return channel can be varied so that a variable flow resistance occurs (corresponding to that provided by the partial barrier 25). In the water model and cells with a corresponding design, the dimensions of the return channel are defined as: depth the depth of the bath in the channel (if the width of channel is small and a working anode forms one or both sides of the channel, the effective channel depth is the true depth minus ACD, since the upward current in the neighboring ACD gap will cause a layer of equivalent thickness at the bottom of the return channel to stagnate. This reduced return channel depth was used in the studies of the water model). Because of the sloping cathode, the depth of the return channel will increase as the bath flows from the upper channel to the lower channel.

width = the horizontal width of the channel vertically

on the flow directions in the channel.

length = the horizontal distance between the upper and lower channel (in the water model this is approximate

equal to BFL).

4) The lower channel completes the bath circulation loop by transporting the bath from the return channel or channels and distributing the bath along the anode edge where the bath enters the ACD gap. The dimensions for this channel are defined similarly to those for that one

upper channel. In most cell designs and in the water model, the lower channel has a well, or trough, at the bottom to collect aluminum metal as it flows down the inclined cathode. Such a well is shown in Figures 1 and 9(a).

In all cases, the bath flow rate, Q, is for the model

and the full-scale cells defined as the total volumetric velocity of the bath flow entering the ACD gap from the lower channel.

In order to simulate a real electrolysis cell and to assess the effect of changing certain specific dimensions or operating conditions, the water model could be changed in different ways. The model was made so that it was possible to simulate adjustable partial barriers (no. 25 in Figure 7). The slope of the cathode and anode in the model was varied from 0" to 15° to determine the effect of cathode slope on bath current in the ACD gap. The ACD gap in the water model was varied from 1 to 5 cm. The gas flow could be varied to simulate different current densities .The fluid flow in the model was observed and measured, using dye injection into the ACD gap and into the return channel.

The design of the anode after the underlying sloping cathode has been demonstrated in both laboratory electrolysis cells and in reports from Kaiser Aluminum and Chemical Corporation to the Department of Energy (final report on DOE Contract No. EY-76-C-03-1257, "Energy Savings Through the Use of an Improved Aluminum Reduction Cell Cathode", 30 November 1977).

Figure 9(b) shows the desired flow of gas bubbles, 8, and the direction of the bath flow, 10, in the ACD. From this figure it can be seen that an excessively high speed of the bath at the cathode surface, in the same direction as at the anode surface, can inhibit the draining of metal from the cathode. Studies of the water model have shown that under certain performance conditions three different undesirable phenomena can occur in ACD; these are: Reversed current occurred when the bath which flowed upwards along the bottom of the anode reversed flow direction and flows downwards along the cathode surface as shown in Figure 9(c). In severe cases of reversed current, as shown in Figure 9(d), the bath forms numerous small current eddies in the ACD and there is no, or very little, supply of fresh bath to the ACD. Persistent anode effects can occur due to insufficient supply of fresh bath in a real electrolytic cell. Air blocking occurred when the removal of the anode gas was too slow and the ACD became saturated with large stagnant gas bubbles. Under such conditions, an insufficient bath flow occurs. In a cell operated under such conditions, the gas bubbles can interrupt the local electrolysis reaction, increase the anode current density in the affected areas of the anode, and possibly lead to increased voltage losses due to anode polarization, which in turn can cause anode effects to take effect.

Excessive bubble thickness occurred under certain conditions and the bubbles extended through the bath and contacted the cathode surface. If this occurs in a cell in operation, it will cause a rapid loss of current efficiency due to the rapid back-reaction between the CO 2 ~ bubbles and the molten aluminum metal on the surface of the cathode.

The phenomena mentioned above and net bath current in the ACD gap and

current in the return channels was studied as a function of cathode slope, ACD, flow resistance in the return channel, and simulated anode current density. While all the observed flow characteristics are in accordance with and controlled by the general hydrodynamic principles described earlier, a quantitative presentation of these has not yet been carried out.

The following general cell specifications were used in the water model and subsequent examples to illustrate the typical application of this invention in the design of improved aluminum reduction cells with reduced energy consumption. In practice, these general cell specifications should be chosen so that they suit a certain application and then the present invention should be used to provide specifications for critical cell parts. Heat balance and profitability calculations have been carried out, and these indicate that "drainage cathode" cells should preferably be operated at higher anode current densities than those typically used in industry today. The performance parameters used in the water model investigations are included in figure 1 for comparison with real cells. The general cell design considered is that shown in Figure 1, with return channels connecting the upper channel in the center to the two outer lower channels on either side of the split VSS anode (each half of the cell corresponds to that shown in Figure 7) .

Figures 10, 11 and 12 illustrate the effect the cathode slope, ACD, and the flow resistance in the return channel, R^, have on the net bath current in the ACD gap at a simulated anode current density of 0.68 A/cm2. In these figures, observed reversed current on the cathode surface is indicated by dotted lines.

In all cases this is associated with significantly reduced net bath current, and becomes more severe as the cathode slope is reduced.

Similar results have been obtained for other simulated current densities. A diagram showing the limits for the selection of parameters, constructed from these data, is shown in Figure 15.

A detailed description of the use of such diagrams for the design of cells that have controlled bath flow is given in the following examples.

The general features of the diagram are as follows. The flow resistance is connected to the width of the return channel for given operating conditions. For each return channel width, a graphical presentation of data from figures corresponding to figures 10, 11 and 12, at constant flow rate, leads

to a relationship between cathode slope and ACD that must be met for a given anode current density in order to achieve a sufficient flow rate in the cell, calculated by methods defined below. This relationship between cathode slope and ACD is represented by the curved lines in Figure 15. For each return channel width, these lines therefore represent a condition that will provide a sufficient supply of aluminum oxide. These lines are limited by

boundary conditions which are defined by undesirable hydrodynamic conditions as described earlier. A region that is prohibited due to excessive bubble thickness limits the operating range at small ACD sizes, i.e. to the left of the parameter diagram (Figure 15). Excessive anode (and cathode) slope, which gives a large immersion depth for the anode, constitutes an upper limit at a slope of approx. 15°. A preferred operating range is shown in figure 15, it lies slightly above the relevant curve for a selected return channel width, so that the ACD is as small as possible and at the same time compatible with the area marked "bubble thickness limited". Operation within this range simultaneously ensures a sufficient supply of alumina, through a sufficiently large Q-value, and prevents excessive bath velocity at the cathode surface (and thereby impeding aluminum draining), as well as excessive bubble thickness and associated loss in current efficiency. The following examples illustrate the construction and use of such a parameter diagram.

Example 1 - Selection of sizes of return channels The bath flow rate, Q, (cm<3>/sec.) must be sufficient

until sufficient alumina is supplied to maintain the electrolysis reaction to prevent anode effects in a cell in operation. At maximum power efficiency, 100%,

is the minimum required flow rate, Q, for an aluminum reduction cell in operation, given by the equation:

where Aweight-% Al2°3 is the difference in weight-% A^O^ in the bath that goes into and out of the ACD gap, or, a alumina. The constant 0.008 is derived from the Faraday equation:

For the sake of simplicity, Q can also be expressed in terms of bath volume per

Sec. per unit anode area. I.e., assuming a minimum anode current density of 0.5 A/cm<2>, and a maximum of 5% depletion of the bath at each pass through the ACD, i.e. Awt%

Al-O-, = 5, a minimum value can be calculated from equation I

on Q = 8 x 10 cm<3>/sec./cm<2> anode area.

Cell function and thermal stability are improved if a uniform Al-jO^ concentration is maintained throughout the bath. For the cells defined above, the calculated minimum bath flow rates are 202, 4000 and 12000 cm<3>/sec. for the water model and example cells I and II, respectively.

While the minimum bath currents, Q, are theoretically sufficient, in practice the values should be exceeded to prevent operational problems (e.g. large anode effects, high effective bath resistance and overvoltages due to large bubble volume in the ACD gap). The operating conditions in the cell will to some extent modify the bath flow (e.g. due to edge formation, crust formation etc.) and can lead to flow rates in the bath that are below the theoretical flow rates. For this reason, a "design factor" of 4 or 5 has been applied to the minimum bath flow values, so that the preferred flow rates are 890, 20,000 and 60,000 cm<3>/sec.

for the water model and example cells I and II, respectively.

Data for the water model show a more reliable and stable bath circulation at these preferred Q values than at the minimal theoretical Q values, while values of approx. 450,

10000 and 30000 cm<3>/sec. considered suitable.

Data from the water model shows that the flow resistance in the return channel is a critical part of this invention. When the return channel is restricted (a result of attempts to maximize the anode area in the cell) the bath current becomes increasingly sensitive to changes in the flow resistance in the return channel. Since there are many cell designs with equivalent properties with respect to flow resistance in the return channel, it is beneficial to construct a simplified hydraulic model to provide general design criteria for the return channel. The hydraulic pressure loss, h^, which is due to flow resistance in the return channel is given by the well-known equation:

where

f^. = Fanning's friction factor

v = velocity

L = length (approximately equal to BFL in the water model)

D eq = equivalent hy Jdraulic diameter

g = gravity of the weight

The friction can be a composite value to reflect differences between the bottom and side surfaces of the channel. For a simple, open channel:

where

w = width of the return channel

h = the depth of the water at any point along

the channel (minus the ACD correction in the water model)

h = hp + x sine <t>

h0 = h at the upper end of the anode (ie minimum anode immersion depth)

x = the distance along the return channel length, from the upper channel = cathode slope

Since the observed velocity in the return channel varies

with the water depth, it is preferable to use volume bath flow which is independent of the varying channel dimensions. The volume bath flow, Q, is given by the velocity times the cross section of the channel, or

WE. Q = v h w, therefore

The Q values used in these equations represent the effective average Q of the channel, which is determined by measuring the time it takes for an injected dye to be transported through the channel.

The hydraulic pressure loss can be written:

where Rf is a geometry term for the flow resistance, which depends on the physical dimensions of the return channel, and K f is a fluid/material property coefficient that is less dependent on physical scale and can be determined in practice:

Since the value of h will vary in a sloping return channel, R is usually calculated as an integral function of h over the channel length L, while the width w generally remains constant. Four of the different flow resistances in the return channels that have been used in the water model studies are reproduced in table 2.

The cell design in this invention is analogous to a pumped liquid circuit where the gas bubbles in the ACD gap perform the pumping action and drive the bath around the ACD gap return channel circuit.

An analogous pump efficiency diagram as shown in Figure 13 is used to describe the circulation characteristics of a "drain cathode" cell. The increase in pump efficiency through curves 1, 2 and 3 reflects the increase in cathode slope and/or the decrease in ACD. The flow resistance increases through slopes a, b and c, they reflect increased length and/or decreasing cross-sectional area of the return channel. The observed flow rates through the ACD are determined

of the intersection of the pump efficiency and flow resistance curves. By placing a vertical line across the pump diagram at the minimum flow rate required to provide sufficient A^O^ for the electrolysis reaction in the ACD, it is possible to form a picture of the pump/flow conditions. The abrupt change in the slope coefficient in some of the pump efficiency curves at low flow rates is associated with reversed flow in the ACD. Table 2 shows the effect the flow resistance in the return channel has on the flow rate in the ACD at a constant cathode slope and ACD.

The low flow rate (Q) shows that a resistance factor of 2.7 x 10 cm is outside the limits of this invention under the stated conditions. This is confirmed by the predominantly reversed flow condition observed in ACD. For the stated conditions, an Rf value is -2-4

2.7 x 10 cm within the limits of this invention,

but it is too large to achieve the preferred flow rate of 890 cm<3>/sec. for the water model, this results in some degree of reversed flow. It should be noted that a slight element of reverse flow can in some cases be beneficial since it can improve,

and do not prevent the drainage of aluminum down the cathode surface.

Scaling up the water model to a commercial aluminum reduction cell provides a number of design options within the scope of this invention. The basic requirements that must be achieved in the production cell are:

1) Sufficient bath current, Q, into the ACD to sustain the electrolysis reaction. The minimum sufficient bath current is given by equation I:

That is for a 100 K A cell with a weight % A12C>3 of 0.2, the smallest bath current is 4000 cm<3>/sec.

2) The hydraulic pressure for the gas-induced pumping action (minus the loss due to flow obstructions in the ACD)

is equal to that of the combined h of the return channel, upper channel and lower channel. For a given slope, ACD, current density and anode length, h developed in the water model and in the full-scale cell should be equal.

3) Design with one or more return channels with the correct dimensions can be used so as to achieve the desired effective flow resistance term, R^. The R^ term for the cell is estimated from that for the model using the equation

where account has been taken of the change in friction factors and total cell flow. If the dimensions of the upper and lower channels are equal to or smaller than those of the return channel, the variable mass flow rate in the upper and lower channels must be included in the calculation. In most anticipated cell designs, and the water model, the dimensions of the upper and lower channels are sufficient so that the small h-^ losses can be neglected compared to the losses in the narrower return channel. The effect of the flow resistance in the ACD is included in the reported data for pump efficiency. Consequently, the Rf link will provide the necessary design criteria for the return channel or the combined effect of several return channels in a full-scale cell. Since a "drain cathode" cell design could include several return channels, n, the effective value for each return channel is given by

XII

where the respective Q's are defined above.

One can now calculate a value for R^ (cell) by summing over all (n) return channels, as given by

and for n equivalent channels,

XIV. R (cell channel) = n<J>Rf (cell),

where R f (cell) can be calculated from the equation

In equation XV, R,(model) is a geometric resistance factor

-3 4

which has a value from 230 x 10 /cm to

0.2 x 10<-3>/cm<4>.

If e.g. the friction factors for the water model and example cells (as indicated in Table 1) are assumed to be equal, the preferred Rf term for each return channel is given by Equation XII, and the width of the channel can be calculated from the respective equations.

Cell example I

Rf (cell channel) = Rf (model) [4(890)/20000]<2 >Rf (cell channel) = 0.032 Rf (model)

w (cell channel) = 20 cm

Cell example II

Rf (cell channel) = Rf (model)[4(890)/60000]<2 >Rf (cell channel) = 0.0035 Rf (model)

w (cell channel) = 52 cm

The return channel widths given above are based on a "design factor" of 5 in the preferred bath flow (Q), and a preferred value R^ (model) = 6.2 x 10 . If the cell operation for cell example II was considered stable enough to reduce the bath flow "design factor" to 3, then the calculated w (cell channel) value would be reduced to 33 cm. An alternative method to reduce the calculated w (cell channel) value to 33 cm, without sacrificing the "design factor", is to increase the anode immersion depth to 8 cm (ie hg = 18 cm). These alternatives demonstrate the utility of the present invention in designing a "drain cathode" cell.

Figure 14 shows the R^ (model) values (as a function of cathode slope and ACD) required to achieve the preferred Q value of 890 cm<3>/sec. for the model. R^ values for other Q values can be extracted from data from the water model which

shown in figures 10, 11 and 12.

Example 2 - Selection of ACD

The cell efficiency is improved by reducing the ACD (and thereby the voltage loss in the bath) without a significant loss in current efficiency. At larger ACD values (eg 4 cm or greater) there is a strong tendency for reverse flow in the ACD gap. If this reversal becomes too large, the Q value will decrease too much. Reduced ACD values promote the desired bath flow through the ACD gap, and higher flow resistances in the return channel can be tolerated. However, at very small ACD values (about 1 cm), a secondary effect between the bubbles and the proximity of a solid carbon surface will delay the desired bath flow (a result of the increased flow resistance in the ACD gap). Figures 10, 11 and 12 show that the maximum bath current does not always occur at the smallest ACD value that one might expect.

The measured maximum bubble thickness decreased from 1.0 cm

to 0.5 cm when the cathode slope increased from 2° to 5°,

with a corresponding increase in net flow rate of the fluid. As the pitch was increased beyond 5°, the bubble thickness remained approximately constant and the bubbles were observed to assume the characteristic shape described earlier and to move more slowly. Changing the ACD was found to have little effect on the observed bubble thickness. If the bubble extends over a large part of the ACD, the current efficiency will be reduced significantly by the back reaction of CO2 with the aluminum metal film on the cathode, and the electrical resistance in the bath will increase greatly. For these reasons, and for practical reasons, the preferred ACD is from 2 to 4 cm, with a more preferred range from 2 to 3 cm.

Example 3 - Choice of cathode slope

In all cases, except when the return channel is very narrow (Figure 10) or the ACD is 4 cm or greater, the bath flow,

Q, exceed the preferred value of 890 cm<3>/sec. at cathode inclinations greater than or equal to 5°. Increasing the cathode slope prevents a large degree of reverse flow,

and reduces the problem of air blockage (large stagnant gas bubbles) in the ACD gap. On the other hand, too steep an angle can break the flow of molten metal on the cathode surface. Back pressure due to constrictions in the return channel can also be prevented by increasing the cathode slope.

However, large cathode slopes can produce large currents

of bath up the slope of the ACD gap, this can prevent the draining of aluminum from the cathode surface and cause practical problems with the large variation in immersion depth of the anode in the bath. When the change in bubble thickness described in example 2 is also taken into account, the preferred cathode inclination is in the range from 5° to 11°,

with the most preferred slope approx. 8°.

Example 4 - Boundary diagram for the selection of cell parameters The construction and use of the previously mentioned diagrams for the cell parameters will now be dealt with in detail. The parameter constraints used to construct this diagram are listed in Table 3.

For practical reasons, and to make the visualization of the cell design easier, Figure 15 indicates flow resistance in the return channel, R^, expressed at corresponding channel widths (assuming a channel length of 121 cm, an anode current density of 0.68 A/cm<2 >, a flow rate in the bath of at least 890 cm<3>/sec., and a minimum channel height of h^ = 10.0 cm). The most preferred set of cell parameters for the water model under the stated conditions is:

cathode slope =7.5°

ACD = 2.5 cm

Rf (return channel) = 6.2 x 10 -3 /cm4

bath current, Q = 890 cm<3>/sec.

The following areas for the critical hydrodynamic flow parameters in the cell are based on: 1) measurement data and observations made during simulation experiments in the water model; 2) the presented analysis of the data; and 3) practical considerations for scaling up to commercial cell size.

A set of parameter ranges is given in Table 4, for the water model which has a BFL of 122 cm, and a BFW of 61 cm. Parameters for a corresponding aluminum reduction cell

in commercial size is given in Table 5.

The criteria in this invention can be applied to several types of cells, such as e.g. those shown in Figures 1 to 7. In particular, the cell may consist of an anode which lies across the width of the cell as a continuous part. An example of this anode is shown in Figure 2 where there is no central vent (2), and consequently the bubble flow length (BFL) is approximately equal to the total anode width. This type of cell design would produce a narrow cell (about 122 cm wide anode for the most preferred case).

Another type of cell would be the one shown in Figure 2, where

the central ventilation (2) is included in the design. The central vent releases the bubbles that have accumulated

clayd in the bath current during passage of the first half of the anode. In this case, BFL is defined as half of the total anode width. Venting of the anode gas is necessary to prevent a large bubble volume from accumulating in the ACD gap, this can increase the voltage drop and decrease the current efficiency. In this type of cell, the total working width of the anode for the most preferred cases is approx. 244 cm, which is a practical cell width.

In a commercial cell, it is desirable to operate at the highest possible anode current density without loss of current efficiency (to minimize capital and labor costs per

kg produced aluminum). At the low ACD values used in a cell with a "drainage cathode", these high current densities will be advantageous for maintaining the heat balance in the cell. At high anode current densities, however, the gas bubbles will accumulate faster and grow larger in the ACD gap than at lower current densities. Accordingly, to counteract the degrading effect of bubble accumulation in ACD, the preferred BFL (anode dimension in the direction of bubble flow) is reduced in proportion to the increase in anode current density. For example, if the anode current density is increased from 1.0 to 2.0 A/cm<2>, the preferred BFL should be reduced from 122 to 61 cm. For an inverted V cathode cell, the preferred design for lower anode current densities is shown in Figure 1.

To increase the preferred overall width of the working anode

at higher current densities (greater than about 1.3 A/cm<2>), the principle of one air hole for releasing the anode gas can be extended to several air holes so that the width of the anode is further increased. The air hole may be present as a gap between neighboring anode masses, or as a row of holes suitably spaced from each other through the anode mass.

Another method is to use a cell of the design shown in Figure 1. In this case, the channel in the center has several purposes, as an air hole for the gas, and as an upper channel that leads the bath coming out of the ACD gap to the return channel or - the channels. In this case, BFL is defined as half the total width of the working anode, this gives a more practical size for the commercial cell. It is understood that the cathode slope shown in figure 1 can be reversed with a central well for collecting metal. In this last case, the upper channels will be located along the outer sides of the cell as shown in Figure 4.

It is also possible to combine these cell designs to

a multi-component cell consisting of many cells connected by the bath and/or metal flow or a single cell compartment containing several neighboring units of the type described above.

For all such cell designs, the dimensions of the return channel or channels are calculated in accordance with the indications in Example 1. The first step is to select the desired bath flow, Q, based on the size of the cell being designed,

and then use the developed relationships to determine the correct flow resistance term, . The equations for the hydraulic pressure loss can then be used under the stated conditions to calculate a set of preferred designs of the return channel. Heat balance, and other criteria regarding cell design or operation, are then used to select one of the hydraulically equivalent designs for the return channel. This final choice of return duct design can now be made with regard to the effect on the critical bath flow.

Example 5 - Test cell with drain cathode

Experimental data from an aluminum reduction cell with inclined TiB2~cathode, carried out at laboratory scale, show the usefulness of

to use a water model to simulate the hydrodynamics of

a commercial aluminum reduction cell. A graphite box with dimensions 61 cm x 122 cm x approx. 46 cm filled with approx.

270 kg cryolite bath, was heated in a closed electric furnace. Once the bath had melted, the experimental cathodes and anodes were lowered through an opening at the top and placed in the cryolite bath as shown in Figure 16. Equipment for automatic supply of A^O^ and lowering of anodes made it possible to carry out continuous electrolysis experiments for up to 3 weeks without any interruption for the replacement of anodes. Under all electrolysis conditions, the output slope (usually horizontal) was

on the lower anode surface after 1 to 5 days of electrolysis transformed into the slope on the cathode surface below. The unevenness on the anode surface shown in Figure 13 is consistent with the bubble behavior observed in the water model. At the starting edge of the anode surface in the water model (corresponding to area A in Figure 16), the bubbles have a very slow and irregular movement which results in a very non-uniform coverage of the anode surface. This will result in an uneven consumption of anode carbon. As the bubbles move upwards along the anode surface, they gain speed and create a more uniform current along the anode surface. Electrolysis under these flow conditions produces a more uniform consumption of anode carbon as shown in area C of Figure 16. Visual observations of the flow patterns in the experimental electrolysis cell are consistent with observations made in the water model.

Example 6 - Data from electrolysis experiments

The cell voltage was measured as a function of current density, ACD value and cathode slope in the electrolysis test cell described in Example 5. Reference voltage for a conventional metal pad cathode was obtained by replacing the inclined TiB2 cathode with a 3 cm deep aluminum pad. in the experimental cell for electrolysis. Figure 17

shows the dramatic reduction in cell voltage noise achieved by replacing an unstable horizontal aluminum pad with a sloping TiB2 cathode in an aluminum reduction cell. In addition to the reduction in the noise in the cell voltage

(i.e. improved cell stability), the experimental cells with a drain cathode showed the following characteristics.

1) The voltage trace for the drain-cathode test cells showed an absence of spikes as shown in Figure 17, caused either by anode effects due to insufficient bath current, or anode-cathode shorts due to turbulence in the metal pad (again a result of bubble thickness and large degree of reversed current). At low ACD values (e.g., less than 2 cm), a conventional control system for the cell voltage (by raising and lowering the anode) required significant operator intervention to conduct an experiment with a metal pad cathode, while experiments with sloping TiB2~cathodes required little or no intervention

by operator.

2) Stable test operation was carried out with a sloping cathode at

ACD values less than 1 cm.

3) The cell's response to changes in current density or ACD was fast and clear. The new stationary cell voltage settled within approx. 1 minute, compared to the 15-30 minute recovery period required for the metal pad cathode. This makes it possible to use simpler and more reliable automation techniques in the cell. 4) Venting of bubbles from the anode surface was mainly along the upper edge of the inclined anode surface.

When an anode with a horizontal surface was placed in the cell with an inclined cathode, the incipient venting of the anode gas bubbles was randomly distributed around all four edges of the anode surface. As electrolysis continued, the anode surface began to take shape following the inclined cathode, and venting of the gas bubbles was concentrated along the upper edge of the inclined anode surface. In experiments where a metal pad cathode was used, only random venting of gas bubbles from the anode surface was observed.

In addition to improved cell stability, the electrolysis experiments show that a significant reduction in the anode-cathode polarization voltage can be achieved by using a drain cathode.

Figure 18 shows the measured anode-cathode polarization voltage as a function of the slope of the TiB2 cathode in the test cell. With the preferred cathode inclination of 8°, the test results show a reduction of approx. 45% in the anode-cathode polarization voltage, or a saving of approx. 0.22 V in the cell voltage in addition to that achieved by reducing the ACD. This advantage of a "drain cathode" cell has not been reported in the literature to date.

The observed decrease in concentration polarization

as a function of increasing bath flow rate (see Figure 18, and also Figure 12 as an example of flow rate as a function of cathode slope) is consistent with electrochemical theory for flowing or stirred systems. In the laminar region, the concentration ion polarization voltage is proportional to the square root of the bath velocity. The measured voltages in Figure 18 do not decrease as quickly with increasing cathode inclination as assumed for laminar flow. This suggests that the conditions in the bath in ACD are either in the transition region or in the turbulent region where the change in concentration polarization voltage is less sensitive to the flow rate in the bath. Data from the water model confirm this hypothesis.

Experimental data for electrolysis in a cell with a tilted TiB2~ cathode compared to a cell with a horizontal metal pad cathode has shown improved cell stability (ie improved current efficiency and reduced anode effects), at reduced ACD values (ie to reduce cell voltage), improved response to changes in ACD and power (i.e. possibility to use simpler and more reliable control automation),

more localized venting of anode gas (ie simpler gas control systems) and reduced anode-cathode polarization voltage. The asymptotic part of the polarization curve at cathode inclinations greater than 5° in figure 18 is in accordance with the preferred range for cathode inclination as suggested by the water model studies.

Claims (18)

1. Drain cathode cell for the electrolytic reduction of alumina to aluminum in a cryolite-based bath containing alumina, characterized in that it includes: a shell having an inner surface lined with refractory materials and carbon materials, so as to define a cathode cavity; a cathode having at least one upper surface containing a hard aluminum wettable refractory material; at least two anodes extending into the cathode cavity, both of which have a lower surface spaced from said upper surface of the cathode at an anode-to-cathode distance of 1 cm to 5 cm, defining an anode-cathode distance ( ACD) and a flow path between anode and cathode where the upper cathode surface and said lower anode surface are inclined from the horizontal plane by from 2° to 15°, so that a lower end and an upper end of the flow path between said anode and said cathode are provided; a device in the cavity to provide during operation a minimum bath flow rate Q(cell) around the at least one substantially horizontal bath circulation circuit to ensure supply of alumina for the electrolysis reaction wherein (where Awt% AI2O3 is the difference in wt% A12O3 in the bath at the entrance and exit of the ACD gap) by providing at least one return flow path to complete the loop, the loop including: (i) said flow path between the anode and the cathode; (ii) at least one lower channel in fluid contact with the lower end of said flow path between the anode and the cathode; (iii) at least one upper channel in fluid contact with said upper end of the flow path between anode and cathode; (iv) at least one return flow channel in fluid contact with said upper channel and said lower channel; the return flow channel, placed between neighboring anodes, has dimensions h, w and L, where h is the bath depth in cm at any given point, w is the width of the return channel in cm, and L is the length of the return channel in cm, and where h , w and L are determined by the relation where Rf is the flow resistance in the return flow channel, and in which the maximum value for Rf (cell channel) is determined by the relations and wherein (X) ranges from 0.2 x 10~<3>/cm<4> to 230 x 10~<3>/cm<4> and where (X) is a geometric resistance factor. 2. Drainage cathode cell according to claim 1, characterized in that said ACD lies in the range from 1.5 to 4.0 cm. 3. Drainage cathode cell according to claim 1, characterized in that said ACD lies in the range from 2.0 to 3.0 cm. 4. Drain cathode cell according to claim 1, characterized in that ACD is approx. 2.5 cm. 5. Drainage cathode cell according to claim 1, characterized in that the upper cathode surface and said lower anode surface form an angle with the horizontal plane of from 5° to 10°. 6. Drainage cathode cell according to claim 1, characterized in that the upper cathode surface and the lower anode surface form an angle with the horizontal plane of from 6° to 8°. 7. Drainage cathode cell according to claim 1, characterized in that the upper cathode surface and the lower anode surface form an angle with the horizontal plane of approx. 8". 8. Drainage cathode cell according to claim 1, characterized in that L(cm) lies in the range from 15 to 300. 9. Drainage cathode cell according to claim 1, characterized in that L(cm) lies in the range of from 30 to 250. 10. Drainage cathode cell according to requirement l.characteristics characterized by L(cm) being in the range of 60 to 200. 11. Drainage cathode cell according to claim 1, characterized in that L(cm) is approx. 122. 12. Drainage cathode cell according to claim 1, characterized in that X(10~<3>/cm<4>) is in the range of from 1.0 to 20. 13 . Drainage cathode cell according to claim 1, characterized in that X(10~<3>/cm<4>) is in the range of from 5.0 to 10. 14 . Drainage cathode cell according to claim 1, characterized in that X(10~<3>/cm<4>) is approx. 6.2. ■ 15. Drain cathode cell according to claim 1, characterized in that the cell includes a large number of return current channels = n, and wherein Rf(cell channel) = n<2>Rf(cell). 16. Drain cathode cell according to claim 1, characterized in that said slope and said ACD are related to each other in such a way that the thickness of bubbles generated in the electrolytic reaction is less than 50% of said ACD. 17. Drainage cathode cell according to any one of claims 1-15. characterized in that a return current channel is arranged between an anode and a nearby inner wall of the cathode cavity. 18. Drain cathode cell according to any one of claims 1-17. characterized in that the aluminium-wettable, hard refractory material is selected from a group comprising titanium diboride, titanium carbide, zirconium diboride, zirconium carbide and mixtures thereof.