• E-mail
• Print
• Comment
• Font Size
• Digg
• del.icio.us
• Discuss article

Long-Term Material Balance of Iron in Aluminum Reduction Cells

Posted on: Friday, 17 March 2006, 06:00 CST

By Simko, Frantisek; Danek, Vladimr; Stas, Milan

The distribution of iron was evaluated using the total material balance of iron in aluminum cells. The data on the Fe^sub 2^O^sub 3^ content in primary and secondary alumina, in aluminum fluoride, and in melted and crushed baths, and the data on the iron content in prebaked anodes, anode butts, and produced aluminum in the period 1998 through 2002 were determined using statistical analysis. The average content for the standard deviation of iron oxide, with respect to the iron in each material, was studied for each year, and a trend was calculated for the entire period. From a statistical point of view, the content of Fe in aluminum systematically increased during the years 1998 through 2002. Conversely, it was observed that the content of Fe^sub 2^O^sub 3^ in the crushed bath decreased during the same period. The crushed bath seems to be a very important flow in the process, since it accumulates most of the iron content. One can expect that a substantial part of the iron is distributed directly into the aluminum and the dry scrubbers. Based on the statistical analysis, the material balance of the iron in the prebaked anode cells resulted in the production of 1 t of aluminum. The total iron material balance (for both input and output materials) in the whole period of investigation totaled -642 g Fe. This value was approximately the same for each year. The higher content of Fe in the output materials is most probably caused by secondary sources of iron, which were not incorporated into the balance.

I. INTRODUCTION

IRON is introduced into aluminum cells in different ways. The main sources of iron in an electrolyte are the raw materials, e.g., alumina, anode materials and AlF^sub 3^; the secondary sources are comprised mostly of step tools, e.g., corroded nipples of anode butts, steel rods in the cathode supply, and other technological operations in which steel tools are used.

The entire amount of iron is distributed between the aluminum and the electrolyte. Most of the iron escapes from the cells with the anode gases, which are emitted from the reduction cells. These effluents consist principally of carbon and alumina particles, gaseous fluorides from the bath, and other gaseous compounds generated during electrolysis. The effluents are adsorbed on alumina in dry scrubbers, in which alumina is used as a medium for sorbing the effluents. The advantages of using alumina as a sorbent include its well-known ability to remove fluoride and also its ability to be returned directly back into the cell as secondary alumina. Iron adsorbed on secondary alumina returns back into the electrolyte, also.[1-4] The presence of iron affects the electrolysis in different ways. It decreases the current efficiency and the metal purity, and it also has a negative influence on the cell operation.[5,6]

In the present work, the complete mass balance of iron in the prebaked anode cells for the period 1998 through 2002 was calculated.

II. RESULTS OF STATISTICAL ANALYSIS

The analyzed materials (flows) are listed below, with their frequently used notations:

(a) primary alumina (alumina entering the dry scrubber);

(b) secondary alumina (alumina entering the cells from the dry scrubber);

(c) AlF^sub 3^;

(d) prebaked anodes;

(e) anode butts;

(f) aluminum;

(g) molten bath; and

(h) crushed bath (material consisting of the anode covering together with the covering remains, and some parts of the solid bath that solidify during the exchange operations of the anodes).

The data on the content of iron, with respect to the iron oxide in all input and output materials, were supplied by Slovalco, Inc. (Ziar nad Hronom. Slovakia). The analytical methods used were as follow: X-ray fluorescence spectroscopy (for the baths), optical emissions spectroscopy (for the primary and secondary alumina. AlF,. and aluminum), and atomic-absorption spectroscopy (for the anodes and anode butts). The results of the iron oxide analysis in the primary and secondary alumina, the AlF3, and the molten and crushed baths, and the results of the iron analysis in the prebaked anodes, the anode butts, and the aluminum, in the period 1998 through 2002, were examined by means of statistical analysis.

The samples for analysis were taken in three different time intervals: almost every day. once a week, and once a month. In order to obtain a unique time axis in the interval <0. t, 1> (for five years: <0, t, 5>), the relative time of the analysis was calculated as the ratio of the day sequence number in the year divided by 365. of the week sequence number divided by 52. and of the month sequence number divided by 12.

For every sample of input and output materials, the average content of Fe^sub 2^O^sub 3^, with respect to iron, and the standard deviation of the median for every year, as well as for the whole investigated period, were calculated. In addition, all trends in the entire time period were also calculated by linear regression. The exception is the content of Fe^sub 2^O^sub 3^ in the secondary alumina, for which the analyses were carried out during the period 2001 through 2002; that time period was too short to calculate the trend for the total period 1998 through 2002. The average content of Fe^sub 2^O^sub 3^ in this material for the missing years was estimated on the basis of the content of P^sub 2^O^sub 5^, from the data used in the material balance of P, as calculated by Danek et al.[7] A similar course for both elements was assumed. The results of the statistical analysis are summarized in Table I.

The content of Fe^sub 2^O^sub 3^ in the primary alumina is shown in Figure 1. The average content of Fe^sub 2^O^sub 3^ in the primary alumina in the entire period studied was (0.0176 0.0064) pet. The differences in the average content of Fe^sub 2^O^sub 3^ in the individual years are caused by the producer.

The content of Fe^sub 2^O^sub 3^ in the secondary alumina is illustrated in Figure 2. The average content of Fe^sub 2^O^sub 3^ in the years 1998 through 2000 was estimated as described earlier. From this calculation, the content of Fe^sub 2^O^sub 3^ in the secondary alumina is estimated to be 0.0317 pet Fe^sub 2^O^sub 3^ during the entire time period; this figure was used in the material balance.

The difference between the content of Fe^sub 2^O^sub 3^ in the primary and secondary alumina in the entire period studied is 0.0141 pet Fe^sub 2^O^sub 3^. This value presents the amount of iron adsorbed in the dry scrubber from the anode gases. It was assumed that, of the entire amount of Fe transported to the dry scrubbers, 95 pet was adsorbed and 5 pet escaped to the chimney as emissions.

The content of Fe^sub 2^O^sub 3^ in the aluminum fluoride is shown in Figure 3. The average content is (0.0168 0.0034) pet in the entire investigated period. However, with regard to the low AlF3 consumption (13.5 kg AlF^sub 3^ per 1000 kg of produced aluminum), this value is not important. The statistically important annual increase of 0.0006 pet/year is given by the producer.

The average Fe^sub 2^O^sub 3^ content in the molten bath throughout the entire period was (0.0184 0.0071) pet (Figure 4). The average Fe^sub 2^O^sub 3^ content levels in the individual years have similar values. The values range from 0.0173 to 0.0194 pet Fe^sub 2^O^sub 3^. The annual decrease of 0.0006 pet/year is considered insignificant from the statistical point of view.

The crushed bath contains essentially a higher content of Fe^sub 2^O^sub 3^. The content of Fe2O3 in this material is shown in Figure 5. The average Fe^sub 2^O^sub 3^ content was (0.1798 0.0401) pet, throughout the entire period. The annual decrease of 0.0095 pet/ year is calculated in light of the standard deviation of this value, which is already significant.

Table I. The Average Content of Iron and the Trend in Iron Content in Individual Analyzed Materials, in Individual Years, and in the Whole Investigated Period

Fig. 1-Content of Fe^sub 2^O^sub 3^ in primary alumina in the period 1998 through 2002. The solid lines are the average content in individual years; the dashed line is the trend during the entire period.

Fig. 2-Content of Fe^sub 2^O^sub 3^ in secondary alumina in the period 1998 through 2002. The solid lines are the average eontent in individual years.

Fig. 3-Content of Fe^sub 2^O^sub 3^ in aluminum fluoride in the period 1998 through 2002. The solid lines are the average content in individual years; the dashed line is the trend during the entire period.

Fig. 4-Content of Fe^sub 2^O^sub 3^ in primary bath in the period 1998 through 2002. The solid lines are the average content in individual years: the dashed line is the trend during the entire period.

Fig. 5-Content of Fe^sub 2^O^sub 3^ in crushed bath in the period 1998 through 2002. The solid lines are the average content in individual years: the dashed line is the trend during the entire period.

The iron content in the aluminum is shown in Figure 6. The average content of iron in the metallic aluminum throughout the entire period was found to be (0.0809 0.0103) pet. The content of iron in the aluminum in the period 1998 through 2001 systematically increases by 0.0038 pet/year. From the statistical point of view, this indicates an undesirable trend. A regular inc\rease in Fe content has been observed in six months out of every calendar year. This condition is probably caused by the the consumption of dust from the pot halls. This occurs once every year in a two- to three- year period. The amount of dust from the pot halls is 200 to 600 tons per year: this material contains an even higher content of Fe. The amount of Fe from this source is not included in the material balance calculation, because of incomplete data.

The mean content of iron in the prebaked anodes for the period 1998 through 2002 is (0.0281 0.0072) pet; as indicated in Figure 7, the iron content substantially increased over that time. On the other hand, the average content of Fe in the anode butts for the entire period is (0.0487 0.0310) pet; as shown in Figure 8, the iron content has a decreasing trend. During the period 2001 through 2002, the difference in the Fe content in the prebaked anodes and in the anode butts is very small. The successive mixing of both materials can cause this effect.

Fig. 6-Content of Fe in aluminum in the period 1998 through 2002. The solid lines are the average content in individual years; the dashed line is the trend during the entire period.

Fig. 7-Content of Fe in prebaked anodes in the period 1998 through 2002. The solid lines are the average content in individual years; the dashed line is the trend during the entire period.

Fig. 8-Content of Fe in anode butts in the period 1998 through 2002. The solid lines are the average content in individual years; the dashed line is the trend during the entire period.

III. DISTRIBUTION OF IRON IN THE PROCESS

The content of iron in the crushed bath is several times larger than in the molten one. The substantially higher Fe^sub 2^O^sub 3^ content in this material could be the result of several factors.

(a) Smaller amounts of iron could be introduced by the secondary alumina and the AlF^sub 3^ during the feeding operation; however, part of these materials could remain in the crust and not go into the bath.

(b) Additional iron comes from the anodes. It is a well-known fact that iron adheres to the carbon particles originating from the consumption of the anodes during electrolysis.[8,9,10] The carbon particles, dispersed in the bath, behave as nucleation sites for iron particles; for reasons of density, the iron particles have a tendency to accumulate as "carbon foam" at the surface of the bath. This material is taken from the cell, along with the anode covering, at the exchange operation of the anode. This mass is recycled into the process. Therefore, the amount of iron increases in the process.

(c) A large amount of iron oxide is introduced by "secondary sources," e.g., from the corrosion of the anode nipples, from the steel rods serving as the cathode current supply, from the iron tools employed during the tapping operation, and in the technological operations of the process of material recycling.

From a statistical point of view, the content of Fe in the aluminum systematically increased during the period 1998 through 2001, by 0.0170 pet Fe. However, in 2002, this value was substantially smaller (0.0032 pet Fe). The opposite effect was observed in the content of Fe in the crushed bath. This fact suggests the similarity of these materials, at least with regard to the distribution of Fe. Part of the iron escapes from the process with the anode gases. The increase in the amount of iron in the secondary alumina, in comparison with the primary alumina, demonstrates this fact. This increase is caused by the adsorption of iron on the alumina, in the dry scrubber. The distribution of iron from the bath into the dry scrubber can occur in several ways:

(a) through the evaporation of gaseous iron compounds from the bath; and

(b) through the entrainment of solid particles (carbon, alumina) by the anode gases (during the feeding and crust-breaking operations).

It is expected that the entrainment of the solid particles is the most probable means by which the impurities are transported from the cell to the dry scrubbers.[1,9,10,11]

It has been shown that the crushed bath plays an important role in the process of iron distribution in the aluminum electrolysis process. Most of the iron is accumulated in this material. Stas and Koniar[12] published similar results for phosphorus as an impurity agent. They described the distribution of phosphorus as occurring in exactly the same manner as presented here. The phosphorus content was concentrated in the covering electrolyte, namely, in the area between the anode and the covering material, as well as in the carbon foam. This conclusion supports the theory that the entrainment of solid particles is the most probable means by which of impurities are transported from the cell to the dry scrubber, during the feeding and crust-breaking operations. Grjotheim and Matiasovsk[13] presented the material balance of iron for the Sderberg's-type cell with the dry scrubber. The content of the iron in the anode gases was twice as high as in our case (for cells with prebaked anodes). In the Sderberg's-type cell, the crust-breaking operations were done more frequently; this resulted in more extensive contamination of the anode gases with the impurities placed in the covering.

IV. MATERIAL BALANCE OF IRON

The material balance of iron was calculated on the basis of the average values of the iron and the iron oxide content, in the input and output materials. The basic information needed for the material balance calculation for one average 230 kA pot consists of the following:

(a) the amount of produced Al (kg);

(b) the volume of exhausting gases (m^sup 3^):

(c) the covering of one anode, consisting of (1) the amount of recycled bath (kg) and (2) the amount of primary alumina (kg);

(d) the number of exchanged anodes per day;

(e) the average mass of the anode (kg);

(f) the amount of consumed alumina (kg);

(g) the amount of consumed AlF^sub 3^ (kg): and

(h) the average anode production per ton of Al (kg).

These data were supplied by Slovalco a.s., Slovakia. The material balance is schematically shown in Figure 9.

In order to relate the material balance to 1000 kg of produced Al, the amount of exhausting gases and the covering of one anode have to be divided by the daily Al production in tons. The consumption of the anodes is given by the number of exchanged anodes multiplied by their average mass.

The mass of the input covering was calculated as the sum of the amount of the crushed bath and of the primary alumina per 1000 kg of the produced aluminum, divided by the daily Al production in tons. The average content of iron in this covering was calculated as the weighted amount of iron in both input materials. The mass of the output covering was the same as the mass of the input covering. The average content of iron in the output covering was identical to the content in the crushed bath.

The amount of the iron in the exhausting gases was estimated from the iron content difference in the primary and secondary alumina. It was assumed that the entire difference in the amount of iron in the primary and secondary alumina was transported by the exhausting gases to the dry scrubbers, where 95 pet was adsorbed and 5 pet was expelled into the atmosphere as emissions.

Fig. 9-Material balance in the 230 kA pot, related to the production of 1000 kg aluminum.

The daily amounts of input and output materials used for one average 230 kA pot, both in the individual years and in the entire investigated period, are given in Table II. The calculated amounts of iron in the input and output materials related to 1000 kg of the produced aluminum, in the individual years, are given in Tables III and IV, respectively.

The total iron material balance (for the input and output materials) in the entire period of investigation is equal to -642 g Fe (Table V).

This value is approximately the same as the values for the individual years. The large amount of Fe in the output materials is most probably due to the secondary sources of iron, which were not included in the balance. Also, the Fe could have been generated by the technological operations, in which iron tools were used. Some of the effect on the results of the mass balance comes from the inaccuracy of the statistical analysis, such as from an incomplete analysis, or from a difference in the time of sampling (per day, per month, per year, etc.).

V. CONCLUSIONS

From the statistical analysis and the material balance of the iron, the approximate distribution of iron in the aluminum production process can be suggested as follows.

1. Iron is introduced into the aluminum cells from raw materials (the alumina and the anode materials), and a substantial amount of iron is introduced by secondary sources.

Table II. Daily Amounts of Input and Output Materials Used for One Average 230 kA Pot in Individual Years and Average Values for the Whole Period

Table III. The Amount of Iron in Input Materials Used for the Production of 1000 kg Aluminum

2. The Fe content in the aluminum systematically increases, during the entire investigated period. The opposite effect was observed in the crushed bath. A dependent relationship between the iron content in the crushed bath and that in the aluminum can be assumed.

3. Additional dependencies may be assumed, including between the content of iron in the crushed bath and in the secondary alumina, because of the entrainment of solid particles by the anode gases. It is suspected that this is the most probable path by which the impurities are transported from the cells to the dry scrubbers.

Table IV. The Amount of Iron in Output Materials Used for the Production of 1000 kg Aluminum

Table V. The Material Balance of Iron in the Aluminum Industry Process

4. It has been shown that the crushed bath is a very important material in the process of iron distribution in the aluminum cell.

ACKNOWLEDGMENT

The Slovak Grant Agencies (VEGA-2/4071/04. APVT-5\1-008104) are acknowledged for their financial support.

REFERENCES

1. J. Thonstad. F. Nordmo. and S. Rolseth: Light Met.. 1978. pp. 463-79.

2. L.C.B. Martins: Light Met., 1987, pp. 315-17.

3. W. Zhang. X. Liu. P. McMaster. and M. Taylor: Light Met., 1996, pp. 405-11.

4. E. Sturm and G. Wedde: Light Met.. 1998, pp. 235-40.

5. P.A. Silli, T. Haarberg, T. Eggen, E. Skybakmoen, and A. Sterten: Light Met., 1994, pp. 195-203.

6. K. Grjotheim, C. Krohn, M. Malinovsk, K. Matiasovsk, and J. Thonstad: Aluminium Eleclrolysis: Fundamentals of the Hall-Hroitlt Process, 2nd ed., Aluminium Verlag, Dsseldorf, Germany, 1982.

7. V. Danek, M. Chrenkov, A. Siln, and F. Simko: Pure Aluminium for the Car Industry: Final Report for Industry, Institute of Inorganic Chemistry SAS, Bratislava, 2002.

8. N. Sillinger and J. Horvath: Light Met., 1990, pp. 369-76.

9. J. Thonstad: Proc. 10th Aluminium Symp., Norway, 1997, pp. 5- 12.

10. J.B. Metson: Proc. 9th Int. Symp. Light Metals Production, Tronclheim, Norway, 1997, pp. 259-64.

11. H.G. Johansen: Thesis, The University of Trondheim, NTH, Trondheim, Norway, 1975.

12. M. Stas and M. Koniar: X. Aluminium Symp., Star Lesn, Slovakia, 1999, pp. 93-97.

13. K. Grjotheim and K. Matiasovsky: Aluminium, 1983, vol. 59, pp. 687-93.

FRANTISEK SIMKO, Scientist, is with the Institute of Inorganic Chemistry SAS, 845 36 Bratislava. Slovakia. Contact e-mail: uachsim@savba.sk VLADIMIR DANEK is retired. MILAN STAS. Metallurgical Engineer, is with Slovalco Inc., 965 48 Ziar and Hronom, Slovakia.

Manuscript submitted January 24, 2005.

Copyright Minerals, Metals & Materials Society Mar 2006

Source: Metallurgical and Materials Transactions; A; Physical Metallurgy and Materials Science

More News in this Category