IMPROVEMENTS RELATING TO THE BAYER PROCESS
This invention relates to improvements to the Bayer process, in particular to a reduction of scale in the settler tank.
BACKGROUND ART
In the Bayer process, alumina is refined from bauxite ores. The process comprises digesting the ore in a caustic solution to extract alumina, clarifying the liquor to remove caustic insoluble red mud material and precipitating alumina crystals as its trihydrate from the clarified liquor. Clarifying the liquor involves separating the solid particles from the liquor by settling and if necessary, filtration. The liquor is fed to a mud settler, or primary settler/thickener where it is treated with a flocculant, and as the mud settles, clarified sodium aluminate solution overflows the top of the settling tank where it is passed to subsequent process steps. The settled solids of the primary settler (red mud) are withdrawn from the bottom of the settler and allowed to pass through a countercurrent washing circuit for recovery of sodium aluminate and soda.
The liquor entering the primary settler is highly
* saturated with alumina. In the course of time, it has been found that alumina scale tends to form first on the walls of the primary settler, then grows upwards, forming a shelf over which the liquor must flow. This shelf of scale blocks the overflow weirs, causing the liquor level to rise. Further upward growth, promoted by the slow liquor flow over the shelf results in a further increase
in liquor levels. Because of this, the primary settler tanks typically eventually become overfull and are taken off-line and mechanically descaled. Such mechanical descalrng operations are costly and limit production capacity. In some instances where significant' scaling problems occur, the effective on-line of the tanks is significantly reduced, by up to 50%. In addition to the need to descale the settler, the presence of scale in the settler is undesirable as it leads to a loss of alumina values from the liquor and reduces flocculation efficiency. It would be desirable to find a means to minimize this formation of alumina scale. DISCLOSURE OP THE INVENTION
According to one aspect, the present invention consists in a method for reducing scale in the Bayer process comprising:
(a) digesting an aluminous ore in a caustic liquor to extract alumina values,
(b) feeding the liquor to a settler tank to clarify the liquor by separating caustic insoluble materials; and
(c) precipitating alumina crystals from the clarified liquor, the process further comprising adding a polysaccharide to the feed of the settler in an amount less than 50 ppm, said amount being sufficient to minimize the formation of scale and lengthen the time between descaling operations in comparison with a feed in the absence of polysaccharide.
Unless the context clearly requires otherwise, throughout the description and the claims, the words 'comprise1, 'comprising', and the like are to be construed in an inclusive sense as opposed to an
exclusive or exhaustive sense; that is to say, in the sense of "including, but not limited to".
Preferably the polysaccharide used is dextran. Suitably a high molecular weight dextran having a molecular weight in excess of 500,000 is used although low density dextrans typically having a molecular weight of from 60,000 to 90,000 or 150,000 to 200,000 can also be used. Other polysaccharides can also be used. Combinations of dextran and starch are expected to be effective and may be more cost effective than the use of dextran alone .
Typically the polysaccharide is used in amounts less than 10 ppm, more preferably 0.5 ppm to 5 ppm, most preferably 1 to 2.5 ppm. The applicants have surprisingly found that even such a low amount of dextran is sufficient to minimize scale. It is thought that higher quantities of dextran are required when it is desired to suppress hydrate precipitation (of the order of 100-150 ppm) . Not wishing to be found by any particular theory, however the applicant believes that although lower concentrations are used in the inventive process, once the polysaccharide is attached to the red mud its local concentration is much higher than the bulk liquor concentration and thus may inhibit hydrate precipitation on the flocculated solids (i.e. seed poisoning as opposed to liquor stabilization) .
The process may also comprise feeding the separated caustic insoluble materials to at least one washer and adding polysaccharide to the washer feed to minimize scaling in the washer.
By using the method of the invention it is possible to substantially lengthen the time between descaling operations in some cases by over 100%.
BEST MODE FOR CARRYING OUT THE INVENTION
A preferred embodiment of the invention will now be described with reference to the accompanying drawing and examples in which:
Figure 1 is a graph of the liquid level in a settler tank over time. EXAMPLES
A number of settler tanks, hereinafter referred to as Fll, F12, F13 , F21, F22, F23, F24, F31, F32 and F33, were introduced over the course of several months into a Bayer process operation to receive digested aluminous ore. Settler tanks Fll, F13 and F32 were each provided with launder scapers to physically descale the tank online during the Bayer process operation. F32 was also provided with a weir descaler. Settler tanks F12 , F22 and F31 were initially used unmodified following which dextran at various concentrations was added to each of the tanks. In this regard, dextran was added to settler tank F12 on 17 October at an initial concentration of 1.3 ppm. Dextran was added to settler tank F22 on 16 August at an initial concentration of 2.6 ppm, the concentration being dropped over time to a minimum value of 1.3 ppm just prior to 9 October when it was increased back to 2.6 ppm. Dextran was added to settler tank F31 on 23 October at an initial concentration of 1.3 ppm. F21 was initially used unmodified and taken off-line for mechanical descaling following which the tank was brought back on-line on 19 August with addition of dextran added
at various concentrations. In this regard, the dextran was initially added to tank F21 at a concentration of 0.7 ppm the concentration being dropped over time to a minimum value of 0.5 ppm just prior to 9 October when it was increased to 1.0 ppm. Settler tanks F23 and F24 were used unmodified although some on-line descaling of F24 was performed during August . Both tanks were then taken off for mechanical descaling. Both these tanks are yet to be brought back on-line. Finally settler tank F33 was used unmodified and taken off-line for mechanical descaling following which the tank was brought back online and again used unmodified with addition of a high molecular weight, highly anionic sodium polyacrylate flocculant .
In each of the tanks where dextran was added, the dextran was used as an aqueous solution.
The results of the operation of each of the settler tanks with respect to the liquor level is show in Table 1 below and in the accompanying drawing (Fig.l) . An increase in the level in the tank is generally indicative of increasing scale in the tank.
TABLE 1 LIQUOR LEVEL IN SETTLER TANK (VOL%)
Date Fll F12 F13 F21 F22 F23 F24 F31 F32 F33
14/4 24 29 13 90 102
29/4 29 37 31 14 103 127
6/5 35 63 17 117 104
8/5 35 20 26
9/5 25 73 17 115 96
13/5 21 73 29 20 98
15/5 22 67 71 30 25 86 99
20/5 28 77 33 27 94 102
27/5 32 61 80 37 34 111
1/6 0
3/6 33
4/6 46 77 39 34 102
9/6 4
12/6 37 52 83 40 37 4 109
18/6 39 1 56 86 78 42 7
25/6 43 4 60 86 42 44 11
2/7 45 9 63 86 81 47 14
8/7 51 8 73 88 54 81 55 11
16/7 56 14 78 56 83 58 14 0
23/7 60 16 81 62 68 14 0
31/7 66 10 93 65 92 76 22 7
7/8 70 14 101 71 91 78 19 9
15/8 18 98 76 93 88 24 13 0
19/8 'l6 0 72 83 87 27 17 4
22/8 19 0 73 91 90 34 18 7
26/8 21 1 73 90 72 36 21 11
29/8 0 20 2 71 68 37 22 11
2/9 4 24 3 69 75 33 22 11
5/9 3 24 4 75 66 35 21 13
9/9 6 26 3 74 75 40 21 17
12/9 5 24 5 70 75 33 21 19
16/9 8 24 2 72 82 34 22 20
3/10 11 32 1 5 76 39 18 28
9/10 11 34 2 8 77 38 21 31
17/10 15 33 2 5 75 45 21 33
24/10 16 35 6 5 76 42 20 37
31/10 16 35 6 7 77 43 27 39
On 6 October the overflow liquor from each of the tanks was analyzed by XRD using both boiled and unboiled samples. Boiling solubilises solid gibbsite. The
difference between boiled and unboiled samples gives a measure of the amount of solid gibbsite in a sample. The overflow from tanks F21 and F22 showed minimal traces of gibbsite solids. The results from the other tanks were varied but generally showed significant concentrations of gibbsite .
Coupons from each of the tanks were also analyzed. Coupons from tanks F21, F22 and F32 were relatively clean with a fine coating of mud-like scale and a small pile of fine settled scale on the horizontal surface. The remaining coupons showed more typical coarse settled scale and a more significant build-up of scale on the vertical surfaces. The coupon from F33 was heavily scaled.
A/TC ratio drop results for each tank were obtained. A/TC is the alumina/total caustic ratio and is calculated from g/1 Al203 divided by g/1 NaOH expressed as Na2C03. The results from tanks F21 and F22 were no less than the results from the other vessels indicating that dextran is not acting to stabilize the liquor. with reference to Figure 1, 0% in the graph refers to a full tank without the presence of any scale. 100% refers to an overfull tank having liquor up to the roof of the launder. This condition requires the tank to be taken off-line and descaled. It can be seen from the table and the figure that in unmodified tanks without the addition of a polysaccharide or the use of mechanical descalers, the liquid levels and hence the amount of scale rapidly increased over time, see for example tanks F33, F23, F24, F21, F22, F31 and F12 prior to the addition of dextran or on-line descaling. Increases in
the liquor level in excess of 20% during a single month are apparent. When dextran was introduced (see F12, F21, F22 and F31) at the quite low levels mentioned above eg 0.5 - 2.6 ppm, the liquor level in the tank was maintained with little or no upward rise and 'thus minimal scale development . It can be seen that liquor level increases were minimized to less than 10% over the course of a single month. With a drop in dextran concentration (see F21 and F22) , a rise in liquid level occurred however the increase was arrested by increasing the dextran concentration to an optimum value of about 1.3 ppm.
It is clear that the use of quite small dosages of dextran minimizes scale development to a significant extent as compared with the unmodified tanks. The results with respect to the use of mechanical scapers were varied. Scale growth in tank F32 was minimal once a - certain liquid level in the tank was obtained. Tanks F13 and Fll both showed increasing liquid levels and it would appear that these tanks are trending towards a certain value at which the results may plateau similarly to F32. In any event it would appear by comparing F21 with F13 , Fll and F32 that whilst mechanical scapers do inhibit scale growth, dosages or dextran at the abovementioned level inhibits scale growth to a greater extent and thus is better at minimizing scale growth.
The examples clearly show the suitability of dextran as an agent for controlling long term scale growth.
It should be appreciated that the examples refer to a preferred embodiment only and the invention applies equally to other polysaccharides . It will further be
appreciated by those skilled in the art that the invention can be embodied in other forms.