THE EFFECT OF QUALITY OF ALUMINA ON THE ENVIRONMENTAL EMISSIONS AT HILLSIDE ALUMINIUM

M. J. Uys

Hillside Aluminium Smelter

P.O. Box 897

Richards Bay 3900

South Africa

ABSTRACT

Hillside Aluminium Smelter consumes approximately 1000 kt of alumina per annum during the production of aluminium. Alumina is supplied from two different sources.

Evolution of fluoride from the reduction cells is approximately 40 kg F/t Al. The air pollution permit limits the fluoride emissions to 1.0 kg F/t Al. The pot gas containing the fluorides is scrubbed in a Gas Treatment Centre (GTC) before being emitted to atmosphere. Alumina is used as scrubbing medium in the GTCs. At present, the fluoride emission from GTC 1 and GTC 2 is higher than from GTC 3 and GTC 4.

The reason for the differences could be due to several factors i.e. plant design, concentration of fluoride in pot gas, plant operation, process control and alumina properties.

The objective of this investigation is to explain the difference between the emissions at the GTCs, focusing on the influence of the two different types of alumina on the process. Alumina from one supplier (Type A) is used on Potline 1, connected to GTC 1 and GTC 2. Alumina from a different supplier (Type B) is used on Potline 2, connected to GTC 3 and GTC 4.

Typical analyses of the two types of alumina are given as well as the effect of alumina properties on fluoride emission. Experience, at Hillside, on the influence of the two types of alumina in the GTCs is discussed in terms of grain size distribution, B.E.T. surface area and loss on ignition.

Some calculations are shown, related to the fluoride content of the alumina in the reactors and the alumina leaving the GTCs. The actual fluoride content is compared to the calculated fluoride content and the comparison is used to show that alumina in GTC 1 and GTC 2 reactors is closer to the adsorption capacity of the alumina. This is one of the main contributing factors to the difference between GTC emissions.

It is recommended that the fresh alumina flow rate at GTC 1 be increased by reducing the charged alumina flow rate.

 

THE EFFECT OF QUALITY OF ALUMINA ON THE ENVIRONMENTAL EMISSIONS AT HILLSIDE ALUMINIUM

M. J. Uys

1.0 INTRODUCTION

Hillside Aluminium Smelter, located in Richards Bay (South Africa), has a capacity of approximately 500 kt of aluminium per annum and consumes approximately 1000 kt of alumina per annum. The smelter was commissioned during 1995 / 1996. The smelter consists of 2 potlines, each with 288 pots.

As with all aluminium smelters, the evolution of fluorides from the reduction process is an important environmental issue. The air pollution permit, granted by the Department of Environmental Affairs to the smelter, limits the total fluoride emissions to 1.0 kg F / t Al. Approximately 30 to 40 kg F / t Al is evolved during the reduction process. Pot gas containing the fluorides should be contained and scrubbed before being emitted into the air. A small part of the pot gas escapes through the pot enclosure directly to atmosphere. The rest is collected and scrubbed in a Gas Treatment Centre (GTC) before being emitted to atmosphere through the stack. The technology is supplied by Procedair.

Alumina, which is also the raw material in the process, is used as a scrubbing medium. The alumina passes through the Gas Treatment Centres on its way from the harbour to the pots. The use of alumina as a scrubbing medium has a twofold function:

  • To act as adsorbent for the removal of fluoride from the pot gas
  • To be used in the reduction process as fluorinated alumina.

At Hillside, alumina is supplied from two different sources. Alumina from one supplier (Type A) is used on Potline 1, connected to GTC 1 and GTC 2. Alumina from a different supplier (Type B) is used on Potline 2, connected to GTC 3 and GTC 4.

Charged alumina, used in the Fume Treatment Centre (FTC) at the Anode Baking Furnace as a scrubbing medium, is also fed to the potlines through GTC 1 and GTC 3. The charged alumina is transported by road tanker from the FTC to the GTC and introduced into the fluorinated alumina stream after the reactors. The flow rate of charged alumina is about 1.88 t/h compared to the total alumina consumption of about 27.6 t/h per GTC.

At present, the fluoride emission at GTC 1 and GTC 2 is higher than GTC 3 and GTC 4. The objective of this investigation is to explain the difference between the emissions at the GTCs, focusing on the influence of the alumina on the process.

2.0 GAS TREATMENT DESCRIPTION

The scrubbing of pot gasses is performed in four GTCs, each serving 144 pots.

The pots, using point feeding, are enclosed and each pot is connected through a duct to the inlet of the GTC. The main part of pot gas that is evolved during the reduction process is collected and scrubbed in a reactor before emitted to atmosphere. The collection efficiency of the GTCs is about 98% and the treatment efficiency is in excess of 99.4%. The emission from the pots directly into the atmosphere is measured at the roof vents of the potline. The emission from the scrubbing process is measured in the stack of the GTCs.

The smelter’s objective is to limit roof vent emission to 0.51 kg F / t Al and stack emission to 0.14 kg F / t Al. This gives a total of 0.65 kg F / t Al, which will ensure that the total emission is always below the statuary 1.0 kg F / t Al.

Figure 1

Gas Collection and Gas Treatment Process

The collected gas is transported through the ducting to the GTC. Each GTC consists of 13 reactors, each with its own filter unit and exhaust fan.

Fresh alumina is split equally between the 13 reactors and injected at the point where the gas enters the reactor. Pot gas is also split equally between the 13 reactors. Contact between the alumina and gas is maximised through the design of the reactor.

The alumina is filtered out of the gas through bag filters and drops down into the filter bottom hopper. Alumina is also recycled from the filter bottom hopper into the reactor to increase contact time with the gas. Pot gas is transported through the ducts by maintaining a negative pressure that is created by the exhaust fans on the outlet side of each filter unit.

Treated gas exits the GTC through the stack. The overflow of alumina from the filter bottom hoppers contains the fluorides and is transported to the pots as the raw material for the reduction process.

3.0 ENVIRONMENTAL EMISSION RESULTS OF GTCs

The average analysis for fluoride in the pot gas is as follows:

Gaseous fluoride: 34.3 kg F / t Al

Particulate fluoride: 6.5 kg F / t Al

Total fluoride: 40.8 kg F / t Al

The total fluoride content in pot gas for two similar smelters is 29.9 kg F / t Al and 31.1 kg F / t Al respectively.

The average stack emission for each GTC is given in Graph 1. GTC 1 and GTC 2 have higher stack emissions than GTC 3 and GTC 4, with GTC 1 and GTC 2 above the specification limit.

Graph 1

12 Month Average Stack Emission per GTC

4.0 REASONS FOR DIFFERENCES IN EMISSIONS

Several factors contribute to the difference between the environmental performance of the GTCs. These factors are summarised in Table 1 below.

Table 1

Factors Influencing the Performance of Gas Treatment Centres

CAUSE

EFFECT

COMMENT

Plant Design
  • Fresh alumina feed rate
 

Reduced fresh alumina feed reduces the ratio of alumina to fluoride with a possible reduction in efficiency

Charged alumina feed into GTC 1 and GTC 3 reduces the fresh alumina feed
Concentration of fluorides in pot gas Higher concentration of fluoride in pot gas reduces the ratio of alumina to fluoride with a possible reduction in efficiency Insufficient measurement data available to conclude that there are significant differences between the concentration of fluorides in the pot gas from different potlines
Plant Operation
  • Stoppage of reactor units

 

  • Stoppage of fresh alumina feed
The pot gas and alumina are distributed to the remaining reactor units with a reduction in efficiency

The recycled alumina is saturated and no treatment takes place after some time

Operating rates and extent of fresh alumina feed stoppages are approximately the same on all GTCs
Process Control
  • Balance of fresh alumina and pot gas between reactors
Unbalanced fresh alumina and pot gas distribution causes unbalance in reactors with a resultant reduction in efficiency No problem if regular cleaning of the distribution box and regular checking and adjustment of pot gas distribution are performed

CAUSE

EFFECT

COMMENT

Alumina Properties Grain size distribution, B.E.T. surface area and LOI have the most influence on GTC operation and fluoride emission The only major difference between the GTCs

 

5.0 TYPICAL QUALITY OF ALUMINA

The typical analyses of the two types of alumina used at Hillside are given in Table 2. The chemical analyses are omitted because they have no influence on the gas treatment process.

Table 2

Analyses of the two Types of Alumina

PROPERTY

UNIT

SPECIFICATION

TYPE A

TYPE B

Grain Size Distribution

  • - 20 m
  • - 45 m
  • + 150 m

%

%

%

< 10

< 10

< 20

0.86 (0.7)

4.2 (1.0)

6.7 (2.8)

2.1 (0.4)

6.0 (1.0)

11.1 (2.1)

B.E.T. Surface Area

m2/g

65 - 85

78 (2.0)

82 (2.0)

Loss on Ignition

(300 – 1000C)

%

< 1

0.7 (0.07)

0.7 (0.02)

Bulk Density

kg/m3

< 1000

975 (7.5)

956 (5.5)

Angle of Repose

< 40

32 (0.5)

33 (0.6)

Flowability

sec

< 100

70 (5.0)

80 (7.0)

Figures are two-year averages with standard deviation in brackets.

  1. EFFECT OF ALUMINA PROPERTIES ON THE PROCESS

Fluoride emissions and GTC processes are influenced by the properties of the alumina. A summary of the effect of the relevant properties on the GTC process and the fluoride emissions is given in Table 3.

Table 3

Summary of the Effect of Alumina Properties on Fluoride Emissions

PARAMETER

EFFECT

Grain Size Distribution

  • - 20 m
  • - 45 m

No known effect on emissions, except the relationship with B.E.T. surface area. Finer material can cause a higher generation of dust in the system resulting in more dust in the atmosphere

Finer material has a negative impact on the process in general:

  • tendency of blockage of fluidisation material
  • tendency to give false indications on high level probes

B.E.T. surface area

Higher B.E.T. surface area gives a higher adsorption capacity of fluoride in alumina and improve adsorption and reaction with HF, resulting in lower fluoride emissions

Loss on Ignition

Higher water content causes more water to enter the bath and more HF to be evolved from the pot, resulting in higher fluoride emissions.

 

 

  1. EXPERIENCE OF THE INFLUENCE OF ALUMINA PROPERTIES AT HILLSIDE

7.1 Grain Size Distribution

The following influences were noticed on the process:

  • Tendency of blockage of fluidisation material – transport air pressures at airlifts for GTC 3 and GTC 4 fluctuate from time to time. The same phenomenon is not observed on GTC 1 and GTC 2. The fluctuations coincide with higher fines content on the fresh alumina analysis.
  • Tendency to give false indications on high level probes – filter hopper high alumina level probes for GTC 3 and GTC 4 give false indications of high level conditions. The same phenomenon is not observed on GTC 1 and GTC 2.

7.2 B.E.T. Surface Area

B.E.T. surface area is a property of the supplied alumina and will determine the fluoride adsorption capacity of the alumina.

7.2.1 Fluoride Content of Alumina in the Process

Fluoride content of alumina in the reactors is an important parameter in controlling the process, since treatment of pot gas will be limited if alumina in a reactor is saturated.

The fluoride content of alumina in the reactors and of alumina leaving the GTC is summarised for each GTC in Table 4. The table compares the calculated values with the equilibrium that exists in each GTC.

Table 4

Calculated and Actual Figures for the Fluoride Content of Alumina at Different Points in the Process

UNIT

GTC 1

GTC 2

GTC 3

GTC 4

Type of alumina

Type A

Type A

Type B

Type B

% F total in reactors – calculated % 2.29 2.13 2.26 2.12
% F total in reactors - actual average

- standard deviation

% 2.35

0.24

2.06

0.22

2.02

0.21

N/A
% F to potlines - calculated % 2.18 2.13 2.16 2.12
% F to potlines - actual average

- standard deviation

% 2.20

0.35

2.11

0.35

2.01

0.23

1.92

0.24

  • Fluoride content of alumina in the reactors and of alumina leaving the GTC is higher in GTC 1 and GTC 3, because the fresh alumina flow rate through the reactors in GTC 1 and GTC 3 is lower than that for GTC 2 and GTC 4. This is due to the charged alumina introduced into GTC 1 and GTC 3.
  • The actual average fluoride content at the reactors in GTC 1 and GTC 2 is fairly close to the calculated fluoride content. The actual fluoride content for GTC 3 and GTC 4 is lower than the calculated fluoride content. This may be due to the fact that potline 1 fluoride evolutions are higher than that of potline 2. As mentioned in Table 1, insufficient measurement data is available to conclude that there are differences between the concentration of fluorides in the pot gas from different potlines. The calculations were performed using average pot evolutions as measured at the inlet to GTC 1, GTC 2 and one measurement at GTC 3. No measurements are available for GTC 4. However, the lower fluoride content of GTC 3 and GTC 4 clearly show that there is a difference in emissions between Potline 1 and Potline 2.

7.2.2 Adsorption Capacity of Alumina

The relationship between the B.E.T. surface area and the gaseous fluoride adsorption capacity of the alumina is as follows:

(1)

The pot gas contains fluoride in the gaseous form (as HF) and in particulate form. This is the adsorption capacity for fluoride in the gaseous form. Since the pot gas also contains some alumina dust particles, the adsorption capacity for total fluoride content will be higher by a factor derived from the ratio of total fluoride to gaseous fluoride in the pot gas.

Fluoride content of the alumina in the reactor should always be below the adsorption capacity of the alumina. In order to relate fluoride content of the alumina in the reactor to the adsorption capacity of the alumina in each GTC, the adsorption capacity of the fresh alumina was calculated for each shipment of alumina delivered during 1997 and 1998, according to equation (1). The relative frequency distribution of these adsorption capacities for Type A and Type B alumina is shown in Graph 2.

The actual reactor fluoride content for each reactor (weekly analysis) during 1997 and 1998 was also obtained. The relative frequency distribution of these values for each GTC is shown in Graph 2. No data is available for GTC 4.

 

Graph 2

Relative Frequency Distributions of the Adsorption Capacity of Type A and Type B Alumina and the Fluoride Content of the Alumina in the Reactors at GTC 1, GTC 2 and GTC 3

The following is observed:

  • There is a significant difference between the adsorption capacities of Type A and Type B alumina.
  • In the case of GTC 1 and GTC 2, using alumina of Type A is a disadvantage as the fluoride content of the alumina in the reactors is closer to the adsorption capacity of type A alumina.
  • At GTC 3 there is a larger margin between fluoride content of the alumina in the reactors and adsorption capacity of type B alumina.

At GTC 4 it can be expected that the fluoride content in the reactors are even lower and no filters will be saturated due to a higher alumina flow rate.

Fluoride emission is higher from GTC 1 and GTC 2 than GTC 3 and GTC 4, because the higher fluoride content of alumina in the reactors together with the lower adsorption capacity of the alumina in the reactors leaves less margin for error on the operation of the GTC.

7.2.3 Loss on Ignition

It is known that a positive relationship exists between Loss on Ignition and fluoride evolution from the pots. No difference exists between Type A and Type B alumina.

  1. CONCLUSION

The following actions can be taken to solve the problem:

REFERENCES

Grjotheim, Kai and Kvande, Halvor (1993). Introduction to Aluminium Electrolysis, 2nd ed., Dusseldorf, Aluminium-Verlag.