US3294656A

US3294656A - Method of producing aluminium

Info

Publication number: US3294656A
Application number: US230855A
Authority: US
Inventors: Schmitt Johannes
Original assignee: Alusuisse Holdings AG
Current assignee: Alcan Holdings Switzerland AG
Priority date: 1961-10-17
Filing date: 1962-10-16
Publication date: 1966-12-27
Anticipated expiration: 1983-12-27
Also published as: GB970798A; DE1181431B; FR1335410A; CH412351A

Description

Dec. 27, 1966 J. SCHMITT 3,294,656

METHOD OF PRODUCING ALUMINIUM Filed Oct. 16, 1962 2 Sheets-Sheet 1 ATTORN y5,

Dec. 27, 1966 J. SCHMITT 3,294,656

METHOD OF PRODUCING ALUMINIUM Filed Oct. 16, 1962 2 Sheets-Sheet 2 l l f l l 1 I ATT NEYS.

BY W fw? M United States Patent 3,294,656 METHOD 0F PRGDUCIN G ALUMINIUM Johannes Schmitt, Rheinfelden, Baden, Germany, assignor to Swiss Aluminium Ltd., Chippis, Switzerland, a jointstock company of Switzerland Filed Oct. 16, 1962, Ser. No. 230,855 Claims priority, application Switzerland, Oct. 17, 1961, 12,030/ 61 4 Claims. (Cl. 204-67) The present invention relates to a process of producing aluminium in an electrolytic furnace from a fused iluoride (cryolite or chiolite) electrolyte containing alumina.

Processes for operating furnaces having one or more pre-baked or self-baking carbon anodes are well-known. `In accordance with the present invention, a set of conditions under which aluminium can be produced at a particularly high current eiciency over long periods of time has been established, I have found that the temperature of the electrolyte should be maintained between 940 and 960 C., the concentration of alumina in the electrolyte should be maintained between 5 and 7% by weight and there should be an excess of between 5 and 7% by weight of aluminium fluoride over that combined in the cryolite as 3NaF.AlF3.

The current eiciency at any moment can be calculated by a newly-developed and rapid method based on the concentration of carbon dioxide and carbon monoxide in the gases escaping from the anodes (the anode gases) at that moment according to the formula:

Momentary -current efciency (percent):

Samples of anode gases were taken through vertical holes in the anodes. The current efficiency over a long period of operation was determined statistically from a number of momentary values.

By the aforesaid method for determining the current efciency, the inuence of the distance of the lower surface of the anode from the separated aluminium (referred to hereinafter as the electrode distance), the electrolyte temperature, the concentration of alumina in the electrolyte and its acidity (the excess of aluminium fluoride over that amount which is combined in the cryolite as 3NaF.AlF3), on the current efciency, as Well as its course between two anode effects has been determined. The influence of one of these variables on the current efliciency while the remaining variables were maintained substantially constant was tested. FIGURES 1 to 4 of the accompanying Edrawings illustrate the results obtained. All precentages are percentages by weight.

FIGURE 1 shows the inuence of the electrode distance (the distance of the lower surface of the anode from the separated aluminium) and the associated furnace Voltage (voltage drop between the anode and cathode current leads outside the electrolytic furnace) on the current efiiciency. The measurements were carried out at a temperature of 970 C, with a cryolite electrolyte which contained 'an excess of AlF3 of about 3% and a concentration -of alumina of about 4.7%. From the curve shown, which was obtained from 143 separate measurements, it surprisingly appears that the electrode distance has a practically immaterial influence on the current ethciency. 0n increasing the electrode distance from 4 to 6 cm., the current eiciency remained practically constant. With a greater distance, it increased slightly.

FIGURE 2 shows the influence of the electrolyte ternperature on the current efficiency. The experiments were carried out with a cryolite electrolyte which contained an excess of A1F3 of about 1.5% and an A1203 content of about 6.7%. The curve, which was obtained from 94 individual measurements, shows that the current efficiency decreased rapidly with increasing temperature to a surprising extent. At 950 C., the current efficiency was found to be about 86.5%, while at 1000 C. it was only 80.5%.

FIGURE 3 shows the inuence of the concentration of alumina in the electrolyte on the current efciency. The measurements were carried out in a cryolite electrolyte at 970 C. with an excess of AlF3 of about 3.3%. From the curve, which was obtained from separate measurements, it can be seen that the current efliciency increased rapidly with increasing A1203 concentration.

FIGURE 4 shows the iniuence of the A1F3 excess in the cryolite electrolyte on the current eiiiciency. The measurements were carried out in an electrolyte at 965 C. which had ian A1202 concentration of about 5.2%. From the curve, which was obtained from 176 separate measurements, it can be seen that the current eiciency first decreased sharply with Ian increasing excess of AlF3 up to about 4%, and then increasedsharply again from about 5%.

I thus found that the current efficiency was influenced to a surprising extent by the electrolyte temperature, the concentration of alumina and the excess of aluminium uoride in the electrolyte, while the electrode distance in contrast did not strongly influence the current eiciency.

The curves shown in FIGURES 1 to 4 are characteristic in their shape for a wide variety of electrolytic furnaces. The absolute value of the current eciency, however, depends upon each individual type of furnace.

It was from these results that the process according to the invention was developed. In contrast to it, conventional processes are usually operated with an electrolyte temperature of from over 960 C. up to about 1000" C., with a concentration of alumina from about 2 to 5% and with an excess of aluminium fluoride from about 0 to 4%.

If the temperature of the electrolyte falls below 940 C., too strong a crust tends to form on the furnace pot, since the solidification temperature of the electrolyte, depending upon its composition (amounts to CaF2, MgF2, as well as impurities such as P205), is generally from 910 to 930 C. Too high a temperature (higher than 960 C.) reduces the current efficiency and must therefore be avoided.

Only within a narrow range of alumina concentrations were comparatively high current eiiiciency values obtained. If the A1203 concentration is .allowed to sink to below 5%, the current eiciency falls substantially. If so much alumina is introduced into the electrolyte when the crust is broken that its alumina-content rises above 7%, theralumina falls to the bottom of the pot and forms a sludge.

The excess of aluminium fluoride should be maintained constant as far as possible between 5 and 7% by frequent adjustment of the electrolyte composition, so that the minimum current eiciency at a 4% excess of AlF3 is avoided. Although the eiciency increases again when the excess of aluminium iluoride falls below 4%, working under these conditions is undesirable, since the walls of the pot may become worn because the alumina becomes too soluble in the electrotyte, and in consequence i-t may not be possible to produce aluminium having a low silicon content. 0n the other hand, an aluminium iluoride excess of above 7% is undesirable because the solubility of A1203 becomes greatly reduced and the furnace may become more difficult to operate.

1n order to keep within the conditions of the process according to the invention, it has been found to be advantageous to break the crust present on the electrolyte at intervals of 1A to 3 hours, preferably every 1 to 2 hours. Breaking of the furnace crust at such short intervals of time is preferably effected by mechanicalmeans,

preferably automatically. The so-called furnace service includes in particular the breaking of the crust and the supply of alumina as Well as aluminium fluoride. To -avoid supersaturation of the electrolyte with A1203, it may be advantageous to break only parts of the crust instead of the whole crust.

Attempts have been made in the past to eliminate the anode effect, in order to avoid the increased voltage which then arises. I have recognized that the anode effect cannot be eliminated, and that to improve the current efficiency even more, it is necessary to cause the anode effect to occur at least once every 48 hours and at the most twice every 24 hours. The anode effect influences the purification of the electrolyte, in that during the anode effect, both gas-forming impurities and also slag and oxides are driven out of the electrolyte. Moreover it brings about a uniform A1203 concentration.

I have also examined the dependence of the anode effect upon the A1203 concentration and on the temperature. Twenty electrolytic furnaces were deprived of current for about half an hour and in this way the electrolyte temperature was reduced by 25 C. from 965 C. to 940 C. After switching it on again, samples of electrolyte were taken and temperature measurements were begun when the current intensity reached 40,000 amps. At half-hour intervals, specimens of electrolyte were with drawn at previously prepare-d positions for determination of the alumina-concentration and the corresponding electrolyte temperature was measured. The aluminacontent d-uring the anode effect was based lon that in the last specimen of electrolyte obtained before the anode effect.

FIGURE 5 shows the dependence of the A1203 concentration during the anode effect on the electrolyte ternperature (the curve was obtained from 51 individual measurements). The occurrence of the anode effect is favored by a reduction in the temperature in that it begins at -a higher A1203 content. For example, at an electrolyte temperature of 995 C., the anode effect occurred only at an A1203 concentration of about 0.6%, while it took place at 960 C. with an A1203 concentration of as much as 2.5%. It was found that the furnace lmust be operated as cold as possible in order that the anode effect should take place with a relatively high content of alumina, which is necessary to achieve a satisfactory current efficiency. l prefer not to wait simply for anode effects, for the electrolyte temperature may increase too much. In order to let the electrolytic furnace work at a temperature between 940 C. and 960 C., therefore, I prefer to break the crust periodically and so throttle the addition o-f A1203 in order to produce an anode effect. lFor that purpose, I allow the concentration of A1203 to drop to -a level between 3 and 2.5% at least once every 48 ho-urs and at the most twice every 24 hours. Also alumina is preferably added mechanically at intervals of 1A to 3 hours.

By the process according to the invention, it is possible to `achieve a current efficiency of to 96%.

What is claimed is:

1. The method of operating -an electrolytic furnace either with prebaked anodes or with selfbaking Sderberg carbon anodes for the production of aluminum by passing direct electric current through a fluoride electrolyte composed mainly of molten cryolite 3NaF.AlF3 and containing an excess of aluminum fluoride AlF3 above that amount which is combined in the cryolite as Well as dissolved alumina Al203, comprising maintaining during a period of at least 12 hours and at most 48 hours the electrolyte temperature between 940 and 960 C., the concentration of alumina between 5 and 7% by weight and of aluminum fluoride excess between 5 and 7% by weight, and preventing the concentration of alumina and of excess aluminum fluoride from dropping below 5% during said period by frequent additions of alumina and aluminum fluoride to the electrolyte.

2. The method according to claim 1, in which at least every 48 .hours and at most every 12 hours, the concentration of alumina is allowed to fall to a level of 3 to 2.5% by weight until an-odic effect occurs.

3. The method according to claim 2, wherein the concentration .of alumina and excess aluminum fluoride is prevented from droppin-g lbelow 5% by weight during said period by frequently breaking the -furnace crust formed on the electrolyte during saidperiod and adding alumina and aluminum fluoride to the electrolyte during said period sufficiently to maintain the concentration of the alumina in the electrolyte between 5 and 7% by weight during said period and the concentration of excess `aluminum fluoride in the electrolyte between 5 and 7% by Weight during said period.

4. 'I'.he method according to claim 2, wherein the concentration of alumina and excess alumin-um fluoride is prevented from dropping below 5% by Weight during said period by frequently breaking the furnace crust formed on the electrolyte during said period at intervals of 1A to 3 hours, and adding alumina and aluminum fluoride to the electrolyte during said period sufficiently to maintain the concentration of the alumina in the electrolyte between 5 and 7% by Weight during said period and the concentration of excess aluminum fluoride in the electrolyte between 5 and 7% by weight during said period.

References Cited bythe Examiner UNITED STATES PATENTS 2,933,440 4/1960 Greenfield 240-'67 3,029,194 4/ 1962 De Varda 204-67 3,034,972 5/ 1962 Lewis 204-67 JOHN H. MACK, Primary Examiner.

H. S. WILLIAMS, `Assistant Examiner.

Claims

1. THE METHOD OF OPERATING AN ELECTROLYTIC FURNACE EITHER WITH PREBAKED ANODES OR WITH SELFBAKING SODERBERG CARBON ANODES FOR THE PRODUCTION OF ALUMINUM BY PASSING DIRECT ELECTRIC CURRENT THROUGH A FLUORIDE ELECTROYTE COMPOSED MAINLY OF MOLTEN CRYOLITE 3NAF.ALF3 AND CONTAINING AN EXCESS OF ALUMINUM FLUORIDE ALF3 ABOVE THAT AMOUNT WHICH IS COMBINED IN THE CRYOLITE AS WELL AS DISSOLVED ALUMINA AL2O3, COMPRISING MAINTAINING DURING A PERIOD OF AT LEAST 12 HOURS AND AT MOST 48 HOURS THE ELECTROLYTE TEMPERATURE BETWEEN 940* AND 960*C., THE CONCENTRATION OF ALUMINA BETWEEN 5 AND 7% BY WEIGHT AND OF ALUMINUM FLUORIDE EXCESS BETWEEN 5 AND 7% BY WEIGHT, AND PREVENTING THE CONCENTRATION OF ALUMINA AND OF EXCESS ALUMINUM FLUORIDE FROM DROPPING BELOW 5% DURING SAID PERIOD BY FREQUENT ADDITIONS OF ALUMINA AND ALUMINUM FLUORIDE TO THE ELECTROLYTE.