WO2009000050A1

WO2009000050A1 - Electrolytic method for controlling the precipitation of alumina

Info

Publication number: WO2009000050A1
Application number: PCT/AU2008/000953
Authority: WO
Inventors: John Besida; Gordon Parkinson; J. David Genders
Original assignee: Alcoa Of Australia Limited
Priority date: 2007-06-27
Filing date: 2008-06-27
Publication date: 2008-12-31
Also published as: AU2008267782A1; BRPI0811708A2; CN101772468A

Abstract

A method for controlling the precipitation of alumina from a Bayer process solution, the method comprising the steps of: applying a potential between a first region comprising a Bayer process liquor and a second region comprising a caustic solution, wherein an ion permeable membrane is provided between the first region and the second region; and causing transfer of an ion across the ion permeable membrane from one region to the other region, wherein the Bayer process liquor is not directed to the second region.

Description

Electrolytic method for controlling the precipitation of alumina

Field of the Invention

The present invention relates to an electrolytic method for controlling the precipitation of alumina from a Bayer process solution.

Background Art

The Bayer process is widely used for the production of alumina from alumina- containing ores such as bauxite. The process involves contacting alumina- containing ores with recycled caustic aluminate solutions at elevated temperatures in a process commonly referred to as digestion. Solids are removed from the resulting slurry and the solution cooled to induce a state of supersaturation.

Alumina is added to the solution as seed to induce precipitation of further aluminium hydroxide therefrom. The precipitated alumina is separated from the caustic aluminate solution (known as spent liquor), with a portion of alumina being recycled to be used as seed and the remainder recovered as product. The remaining caustic aluminate solution is recycled for further digestion of alumina- containing ore.

The precipitation reaction can be generally represented by the following chemical equation with reference to the precipitation of aluminium hydroxide. A similar equation may be prepared for the precipitation of aluminium oxyhydroxide:

AI(OH)₄ ^" (aq) + Na⁺ (aq) ► AI(OH)₃ (S) + OH^" (aq) + Na⁺ (aq)

As the precipitation reaction proceeds, the A/TC ratio of the liquor falls from about 0.7 to about 0.4 (where A represents the alumina concentration, expressed as gl_^"1 of AI₂O₃, and TC represents total caustic concentration, [NaOH] + [NaAI(OH)₄], expressed as gl_^'1 of sodium carbonate). At the lower value of A/TC, the rate of precipitation slows substantially due to a decrease in the level of supersaturation, and an increase in the level of "free caustic" in the liquor, as the system approaches equilibrium.

It is known that the TC and TA (where TA represents total alkali concentration, [NaOH] + [NaAI(OH)₄] + [Na₂CO₃] , expressed as gl_^"1 of sodium carbonate) of Bayer process solutions affects the solubility of boehmite and gibbsite in those solutions in a number of ways.

Generally, more than half of the alumina stays dissolved in solution, to be recycled through the digestion circuit of the plant. In principle, if some of the hydroxide formed during precipitation could be removed, the A/TC ratio of the liquor would increase and the equilibrium of the above reaction would be shifted to the right favouring more precipitation of alumina. Further, it is believed that supersaturation could also be induced, and controlled, by reducing the level of caustic in Bayer liquor with the benefit of achieving an increase in yield beyond that which is attainable under current practices.

Liquor carbonation is a technique used in the alumina industry to convert hydroxide to carbonate, and has been used to increase the precipitation yield of alumina. However, apart from producing alumina of inferior quality compared to current practices, liquor carbonation necessitates the excessive purchase cost of lime which is required to regenerate caustic from sodium carbonate. Further, the recausticisation step is inefficient and does not result in complete regeneration of the caustic.

Based on alumina solubility alone, the removal of sodium ions in conjunction with the neutralisation of hydroxide should produce a greater increase in precipitation yield than hydroxide neutralisation by carbonation. This is because the former leads to a reduction in both TC and TA whereas the latter leads to a reduction in TC but TA remains constant (where TC and TA represent the total caustic concentration and the total alkali concentration respectively, both expressed as gL^'1 sodium carbonate). For a given value of TC, alumina is more soluble in solutions of higher TA. The preceding discussion of the background to the invention is intended to facilitate an understanding of the present invention. However, it should be appreciated that the discussion is not an acknowledgement or admission that any of the material referred to was part of the common general knowledge in Australia, or anywhere else, as at the priority date of the application.

Disclosure of the Invention

Those skilled in the art will appreciate that the invention described herein is susceptible to variations and modifications other than those specifically described. It is to be understood that the invention includes all such variations and modifications. The invention also includes all of the steps, features, compositions and compounds referred to or indicated in the specification, individually or collectively and any and all combinations or any two or more of the steps or features.

The present invention is not to be limited in scope by the specific embodiments described herein, which are intended for the purpose of exemplification only. Functionally equivalent products, compositions and methods are clearly within the scope of the invention as described herein.

The entire disclosures of all publications (including patents, patent applications, journal articles, laboratory manuals, books, or other documents) cited herein are hereby incorporated by reference.

In accordance with the present invention, there is provided a method for controlling the precipitation of alumina from a Bayer process solution, the method comprising the steps of:

applying a potential between a first region comprising a Bayer process liquor and a second region comprising a caustic solution, wherein an ion permeable membrane is provided between the first region and the second region; and causing transfer of an ion across the ion permeable membrane from one region to the other region,

wherein the Bayer process liquor is not directed to the second region.

Advantageously, by not directing the Bayer process liquor to the second region, a relatively pure caustic stream may be produced from the second region. Such a stream could be used within the Bayer circuit for, for example, washing bauxite (to extract impurities) or any other application where clean caustic is useful. In addition, the Bayer process liquor, after a precipitation step may be used elsewhere in the Bayer circuit. For example, it would have a lower TC/TA content and a lower alumina content than normal spent liquor which should enable more efficient causticisation with lime.

It will be appreciated that more than two regions may be provided and that more than one ion permeable membrane may be provided.

applying a potential between a first region comprising a Bayer process liquor and a second region comprising a caustic solution, wherein an ion permeable membrane is provided between the first region and the second region; and

causing transfer of an ion across the ion permeable membrane from one region to the other region,

wherein the caustic solution has a maximum alumina concentration of 20 gl_^"1 as AI₂O₃.

A lower alumina content in the caustic solution may be advantageous for bauxite washing as this wash stream would not be recycled back to the Bayer liquor thereby reducing alumina losses. It will be appreciated that more than two regions may be provided and that more than one ion permeable membrane may be provided.

wherein there is provided at least a third region and at least two ion permeable membranes.

As used herein the term "alumina" shall be taken to include, without limitation, any form of aluminium hydroxide, aluminium oxyhydroxide or aluminium oxide.

Preferably, the Bayer process liquor is not directed to the second region.

Where there is provided more than one ion permeable membrane, the ion permeable membranes will preferably be substantially coplanar such that adjacent ion permeable membranes will preferably permit the transfer of oppositely charged ions.

In one form of the invention, there is provided at least one anion permeable membrane and at least one cation permeable membrane.

Where there is provided at least one anion permeable membrane and at least one cation permeable membrane, the plurality of membranes may comprise an electrodialysis unit. In a second form of the invention, there is provided at least one bipolar membrane and at least one cation permeable membrane.

Where there is provided at least one bipolar membrane and at least one cation permeable membrane, the plurality of membranes may comprise an electrodialysis unit.

Where there is provided one ion permeable membrane, the ion permeable membrane is preferably a cation permeable membrane and the ion is a cation. Preferably, the cation is a sodium cation.

It will be appreciated that the transfer of the cation from one region to another region will encompass the transfer of more than one cation from the first region to the second region.

Preferably, one region is provided with an anode and another region is provided with a cathode.

It will be appreciated that the transfer of the cation from one region to another region will either be accompanied by a concomitant neutralisation of hydroxide ions within the Bayer process liquor and generation of hydroxide ion in the second region or, in the case of an electrodialysis unit, the transfer of hydroxide from one region to another region, and in the opposite direction to the cation transfer to maintain solution charge balance.

The present invention offers distinct advantages over methods employing carbonation to reduce hydroxide concentrations in Bayer process solutions, as carbonation reduces TC without affecting TA, but the present invention reduces both the TC and TA of the Bayer process liquor. Further, the present invention increases the A/TC of the Bayer process liquor, thereby increasing the precipitation efficiency of alumina.

The present invention offers distinct advantages over methods employing solvent extraction to extract species in order to control the precipitation of alumina in Bayer liquors as the liquor does not need to be contacted with an organic solvent which could result in adverse reactions between constituents of organic phase and constituents of plant liquor. Further, it is not necessary to employ a separate phase separation stage nor a subsequent stripping stage to recover soda from the organic phase as caustic.

Preferably, the method comprises the further step of:

precipitation of alumina in the Bayer process liquor.

Alumina is more soluble in alkaline solutions than in water and advantageously, the reduction of sodium ion concentration in the Bayer process liquor can increase precipitation of alumina.

Advantageously, the method of the present invention may be utilised to control the form of precipitated alumina and influence the formation of forms such as boehmite, gibbsite, bayerite, doyleite and nordstrandite. It will be appreciated that the alumina may be a mixture of any of the preceding forms.

Advantageously, by removing soda and by precipitating alumina, any impurities in the spent liquor are concentrated which may make the stream more amendable to impurity removal techniques such as liquor chilling (to remove carbonate and sulfate as a double salt of sodium).

Preferably, the method comprises the further step of:

seeding the Bayer process liquor with alumina.

In one form of the invention, the step of:

seeding the Bayer process liquor with alumina

is conducted prior to the step of: applying a potential between a first region comprising a Bayer process liquor and a second region comprising a caustic solution, wherein an ion permeable membrane is provided between the first region and the second region

In a second form of the invention, the step of:

seeding the Bayer process liquor with alumina

is conducted after the step of:

applying a potential between a first region comprising a Bayer process liquor and a second region comprising a caustic solution, wherein an ion permeable membrane is provided between the first region and the second region;

It will be appreciated that the optimal seeding rate will depend on many factors, including the seed and liquor properties and the design of the precipitation circuit, and may be greater than 50 gL^'1 and preferably, in the range of 50 to 1300 gL^"1.

Preferably, the step of:

causing transfer of an ion across the ion permeable membrane from one region to the other region;

is conducted at a temperature which is below that used to digest the bauxite.

Preferably, the step of:

precipitation of alumina in the Bayer process liquor.

is conducted at temperatures up to the boiling point of the liquor at that pressure.

Advantageously, the present invention can negate the need to reduce the temperature of a Bayer process liquor to encourage supersaturation. It is known that precipitation rates decrease with temperature. In a gibbsite precipitation circuit, precipitation commences at about 90 ⁰C and ends at about 60 °C at the completion of the precipitation phase. Without being limited by theory, it is believed that the method of the present invention may permit precipitation of alumina at temperatures as high as the boiling point of the liquor at that pressure.

The present invention may be utilised to increase precipitation yields beyond current limits without initially increasing TC in digestion. It may further provide means of inducing supersaturation without appreciable liquor cooling.

Without being limited by theory, it is believed that a sodium/aluminate ion pair exists on or near the surface of precipitated alumina and hinders further deposition of alumina onto the surface. Advantageously, removing sodium from the Bayer process solution may increase alumina precipitation.

Importantly, the present invention does not advocate a measurable reduction in solution pH. Bayer liquor pH is above measurable limits (>14) and it has been discovered that a significant increase in precipitation yield can be obtained by instigating a decrease in sodium ion concentration by present method invention, whereby liquor pH is still kept well above a value of 14.

It will be appreciated that the ion permeable membrane should be substantially resistant to corrosion or degradation under the electrolytic conditions.

It will be appreciated that the choice of ion permeable membrane will be dependant on many factors including the selectivity of ion transport, including the selectivity of sodium ion transport. Further factors include the conductivity and rate of ion transport, the mechanical, dimensional and chemical stability, resistance to fouling and poisoning and membrane lifetime.

In specific forms of the invention, the cation permeable membranes may comprise perfluorinated polymers such as a sulfonated tetrafluorethylene copolymer, carboxylate polymer, polystyrene based polymer, divinylbenzene polymer, or sodium conducting ceramics such as beta-alumina or combinations thereof. In highly specific forms of the invention, the cation permeable membrane is a Nafion 115, Nafion 117, Nafion 324, Nation 440, Nation 350, Nation 900 series, Fumatech FKB, Fumatech FKL membrane, Astom CMB or Astom CMX membrane.

Perfluorinated membranes are known to have a high resistance to chemical attack under conditions of high pH. The stability and favourable physical properties are believed to be due to the substantially inert and strong backbone of the polymer which contains regular side chains ending with ionic groups. The choice of the ionic groups is important as they affect interactions with the migrating ions, the pK_a of the ion exchange polymer, the solvation of the polymer and the nature and extent of interactions between the fixed ionic groups.

In the case of an electrodialysis unit, the anion permeable membrane is preferably a Neosepta AHA membrane or a Fumatech FAP membrane.

It will be appreciated that the electrode material should exhibit high conductivity and low electrical resistance and be substantially resistant to corrosion under the electrolytic conditions. Bayer liquor is highly caustic but H⁺ is produced at the anode. It will be appreciated that choice of electrode material will be within the ability and knowledge of the skilled addressee. Since Bayer liquor contains anions such as fluoride, sulphate etc. the production of hydrofluoric acid, sulfuric acid etc. may occur at the interface between anode and solution (even though the solution is highly caustic). Suitable anode materials include platinum coated niobium, platinum coated titanium or Monel.

It will be appreciated that base only is produced at the cathode so the choice of cathode material may be wider than anode material. Suitable cathodes include stainless steel or a gas diffusion electrode (oxygen depolarized cathode).

It will be appreciated that the current density must be controlled as increasing the current density will increase the rate of product formation but it will also increase the energy consumption. For higher current densities, less membrane area may be required for a given quantity of caustic extracted. For systems employing one cation exchange membrane, the preferred current density may be between 20 mA/cm² and 600 mA/cm². More preferably, the current density is between 150 mA/cm² and 350 mA/cm².

Whilst it is advantageous to have the caustic concentration of the caustic solution as high as possible, if it is too high, the current efficiency may be compromised due to back diffusion of hydroxide ions from the caustic solution to the Bayer process liquor. The caustic solution may be sourced from the Bayer circuit. It will be appreciated that where the caustic solution is sourced from the Bayer circuit, the solution should have a caustic concentration below that of the Bayer process liquor. Non-limiting examples include Bayer lake water or condensate.

Preferably, the caustic concentration is not greater than about 8M NaOH or 25% NaOH. It will be appreciated that if the caustic concentration is too low then the current density may drop due to lower conductivity.

In one form of the invention, the step of:

applying a potential between a first region comprising a Bayer process liquor and a second region comprising a caustic solution, wherein an ion permeable membrane is provided between the first region and the second region,

is performed in a Bayer process side stream.

In one form of the invention, where the Bayer process includes the steps:

digestion of bauxite with caustic solution;

liquid-solid separation to provide a residue and a Bayer liquor; and

precipitation of alumina from the Bayer liquor;

the steps of: applying a potential between a first region comprising a Bayer process liquor and a second region comprising a caustic solution, wherein an ion permeable membrane is provided between the first region and the second region; and

the steps of:

causing transfer of an ion across the ion permeable membrane from one region to the other region.

are conducted prior to the step of:

precipitation of alumina from the Bayer liquor.

In a second form of the invention, where the Bayer process includes the steps:

digestion of bauxite with caustic solution;

liquid-solid separation to provide a residue and a Bayer liquor; and

precipitation of alumina from the Bayer liquor; and

the steps of:

applying a potential between a first region comprising a Bayer process liquor and a second region comprising a caustic solution, wherein an ion permeabJe membrane is provided between the first region and the second region; and

are conducted after the step of:

precipitation of alumina from the Bayer liquor. Where the steps of:

are conducted after the step of:

precipitation of alumina from the Bayer liquor,

the method may comprise the further step of:

precipitation of further alumina from the Bayer process liquor.

In a third form of the invention, where the Bayer process includes the steps:

digestion of bauxite with caustic solution;

liquid-solid separation to provide a residue and a Bayer liquor; and

precipitation of alumina from the Bayer liquor; and

the steps of:

causing transfer of an ion across the ion permeable membrane from one region to the other region. are conducted during the step of:

precipitation of alumina from the Bayer liquor.

The method of the present invention may be performed as a batch process wherein the first region is provided in the form of a first compartment and the second region is provided in the form of a second compartment and the ion permeable membrane is provided between the first compartment and the second compartment. The Bayer process liquor is introduced into the first compartment and the caustic solution is introduced into the second compartment and a potential is applied between the first compartment and the second compartment for a set period of time, after which the Bayer process liquor, depleted in sodium ions and in hydroxide ions is removed from the first compartment and the caustic solution with an increased sodium hydroxide concentration is removed from the second compartment.

Alternatively, the method of the present invention may be performed as a continuous process wherein the first region is provided in the form of a first compartment and the second region is provided in the form of a second compartment and the ion permeable membrane is provided between the first compartment and the second compartment. Bayer process liquor is continuously introduced into the first compartment and caustic solution is continuously introduced into the second compartment with a potential continuously applied between the first compartment and the second compartment. Treated Bayer process liquor, depleted in sodium ions and in hydroxide ions is continuously removed from the first compartment and caustic sofutibn with an increased sodium hydroxide concentration is continuously removed from the secopd compartment.

Alternatively still, the method of the present invention may be performed as a continuous process with many compartments in a cell with adjacent compartments being alternately separated by cation permeable membranes and anion permeable membranes. Every second region contains a feed solution of Bayer process liquor and instead of hydroxide being neutralized by production of protons at the anode, it is removed from the feed solution through an anionic membrane to form relatively pure caustic (sodium ions come in from the opposite side via a cationic membrane). The method is believed to consume less energy than electrolysis with a single ion permeable membrane because the amount of water that is electrolysed to form protons and hydroxide, with concomitant formation of hydrogen and oxygen, is minimized. Optionally, the arrangement could include bipolar membranes in place of anion permeable membranes. Bipolar membranes split water directly, to produce hydroxide ions and protons, with no hydrogen or oxygen formation.

Brief Description of the Drawings

The present invention will now be described, by way of example only, with reference to eight embodiments thereof, and the accompanying drawings, in which:-

Figure 1 a is a schematic flow sheet of the Bayer circuit;

Figure 1b is a schematic flow sheet showing how a method in accordance with a first embodiment may be utilised in the Bayer circuit;

Figure 2a is a schematic representation of an electrochemical cell in accordance with a second embodiment of the present invention;

Figure 2b is a schematic representation of an electrochemical cell in accordance with a third embodiment of the present invention;

Figure 3 is a graph showing the effect of caustic strength on current efficiency

Figure 4 is a graph showing the change in current efficiency when operating the cell at a higher A/TC ratio;

Figure 5 is a graph showing the change in iron concentration in the anolyte for various anodes; and Figure 6 is a graph showing the change in nickel concentration in the anolyte for various anodes.

Figure 7 is a schematic representation of an electrochemical cell in accordance with a fourth embodiment of the present invention;

Figure 8 is a schematic representation of an electrochemical cell in accordance with a fifth embodiment of the present invention;

Figure 9 is a schematic representation of an electrochemical cell in accordance with a sixth embodiment of the present invention;

Figure 10 is a graph showing the effect of temperature on voltage and current density using a gas diffusion electrode as an anode;

Figure 11 is a graph showing the effect of temperature on voltage and current density using a gas diffusion electrode as a cathode;

Figure 12 is a graph showing the effect of temperature on voltage and current density using conventional flat plate electrodes;

Figure 13 is a graph showing the effect of a gas diffusion electrode, acting as either an anode or cathode, on voltage and current density as a comparison with conventional electrodes;

Figure 14 is a schematic representation of an electrochemical cell in accordance with a seventh embodiment of the present invention;

Figure 15 is a graph showing the effect of bipolar membranes;

Figure 16 is a graph showing the amount of Al transport across the FKB membrane;

Figure 17 is a schematic representation of an electrochemical cell in accordance with an eighth embodiment of the present invention; Figure 18 is a graph showing the concentration profile of spent Bayer liquor during electrodialysis;

Figure 19 is a graph showing the effect of current density on voltage and charge; and

Figure 20 is a graph showing the effect of current density on voltage and charge for a constant catholyte caustic concentration of 9 % w/w.

Best Mode(s) for Carrying Out the Invention

The invention focuses on the control of alumina precipitation in the Bayer process by transfer of sodium ions from a Bayer process solution through an ion permeable membrane under the influence of a potential gradient. By careful manipulation of the extraction conditions, the precipitation of alumina from aluminate solutions may be controlled.

Figure 1a shows a schematic flow sheet of the Bayer process circuit for a refinery using a single digestion circuit comprising the steps of:

digestion 12 of bauxite 14 in a caustic solution;

liquid-solid separation 16 of the mixture to residue 18 and liquor 20;

alumina precipitation 22 from the liquor 20;

separation of alumina 22 and liquor 24; and

recycling liquor 24 to digestion 12.

In accordance with a first embodiment of the present invention and best seen in Figure 1b, there is further provided an electrochemical cell 26 comprising an anolyte compartment 28 and a catholyte compartment 30 separated by a cation permeable membrane 32 wherein the liquor 24 is pumped through anolyte compartment 28 and a caustic solution 33 is pumped through the catholyte compartment 30. A potential is applied across the electrochemical cell 26 and sodium ions transported across the membrane 32 to the catholyte compartment 30. Concurrently, a proton is produced at the anode from the oxidation of water neutralised hydroxide in the anolyte.

The treated liquor 34 is removed and may be seeded to induce alumina precipitation 36. It will be appreciated that the spent liquor after removal of the precipitated alumina 36 may be further treated before returning to digestion as shown by the dotted line in Figure 1b..

In accordance with a second embodiment of the present invention, best seen in Figure 2a, there is further provided a electrochemical cell 37 comprising a plurality of alternating anolyte compartments 38 and catholyte compartments 40, each compartment alternately separated by a cation permeable membrane 42 and an anion permeable membrane 44 wherein the liquor 24 is pumped through the anolyte compartments 38 and a caustic solution 33 is pumped through the catholyte compartments 40. A potential is applied across the electrochemical cell

37 and sodium ions transported across the cation permeable membrane 42 to the catholyte compartments 40. Concurrently, hydroxide ions are transported across the anion permeable membrane 44 to the catholyte compartment 40. Overall, this process depletes the liquor 24 in causticity and increases the causticity of the catholyte.

The treated liquor 34 is pumped out of the anolyte compartments 38 and may be seeded to induce alumina precipitation 36. The solution 45 exiting the catholyte compartments 40 has increased causticity.

In accordance with a third embodiment of the present invention, best seen in Figure 2b, there is further provided a electrochemical cell 46 comprising a plurality of alternating anolyte compartments 48 and catholyte compartments 50, each compartment alternately separated by a cation permeable membrane 52 and a bipolar membrane 54. Liquor 24 is pumped through the anolyte compartments 48 and a caustic solution 33 is pumped through the catholyte compartments 50. A potential is applied across the electrochemical cell 46 and sodium ions transported across the cation permeable membrane 52 to the catholyte compartments 50. Concurrently, a proton produced at the anode from the oxidation of water neutralises hydroxide in the anolyte.

The treated liquor 34 is pumped out of the anolyte compartments 48 and may be seeded to induce alumina precipitation 36. The solution 56 exiting the catholyte compartments 50 has increased causticity.

The following Examples serve to more fully describe the manner of using the above-described invention, as well as to set forth the best modes contemplated for carrying out various aspects of the invention. It is understood that these Examples in no way serve to limit the true scope of this invention, but rather are presented for illustrative purposes.

Unless stated otherwise, all experiments utilised a divided flow cell consisting of anolyte and catholyte compartments separated by a Nafion^® cation membrane available from Du Pont. Nafion is a sulfonated tetrafluorethylene copolymer. Conventional methods of determining molecular weights of Nafion membranes, such as light scattering and gel permeation chromatography, are not applicable because Nafion is insoluble Instead, the equivalent weight (EW) and material thickness are used to describe most commercially available membranes. The EW is defined as the weight of Nafion per mole of sulfonic acid group. For example, Nafion 117 represents 1100 g EW + 0.007 in thickness.

The anolyte (Bayer spent liquor from the Applicant's refinery at Kwinana, Western Australia) was pumped through the anolyte compartment whilst the catholyte (caustic) was pumped through the catholyte compartment. The catholyte was a synthetic solution prepared from sodium or potassium hydroxide. In a plant, it is envisaged that a portion of the catholyte would be bled from the cell, sent for mixing with spent liquor prior to digestion, and replaced by lake water or condensate to reduce the causticity before recycling the catholyte back through the cell. With a potential applied across the cell, sodium ions were transported across the membrane to the catholyte. Concurrently, a proton produced at the anode from the oxidation of water neutralised hydroxide in the anolyte. Anode H₂O 8^" + VzO₂ (g) + 2H⁺ (aq)

Anolyte H⁺ (aq) + OhT (aq) ^H₂O

Cathode H₂O + e^{" ►} OH^" (aq) + ¹/₂H₂ (g)

Catholyte Na⁺ (aq) + OH^' (aq) ►NaOH (aq)

The above equations clearly show how hydroxide, in the anolyte, is neutralised by the proton formed at the anode with the concomitant formation of hydroxide ions at the cathode. The combination of these reactions removes caustic from the anolyte and supersaturates the anolyte with respect to alumina solubility. The anolyte can then be contacted with alumina seed for precipitation of alumina. The catholyte, which contains increased levels of caustic, may be re-used in the Bayer circuit. Low concentration catholyte could be used for bauxite washing. Higher concentration catholyte could be either added to spent liquor directly, before routing to digestion, or combined on its own with bauxite during the milling stage prior to digestion.

Gibbsite was used as seed for all experiments involving precipitation and these were conducted as batches using polypropylene bottles of 250 ml_ capacity positioned in a rotating water bath. Unless stated otherwise, 10 g of seed was used per 100 ml_ of liquor for precipitation experiments.

Experimental conditions and solution compositions were intended to replicate those conditions and compositions present at the backend of the alumina precipitation circuit of refineries.

At the end of the contact period, precipitated alumina was collected by filtration, washed with hot water, dried in an oven at 105 ⁰C and weighed. The aqueous filtrate was stabilised by the addition of sodium gluconate to prevent gibbsite precipitation from solution upon liquor cooling to room temperature and analysed for TC and alumina content by titration and ICP. All solid samples from the precipitation experiments were analysed by XRD and found to consist of gibbsite only.

The entire electrolysis setup was contained within a fume hood to facilitate proper venting of the hydrogen and oxygen produced at the electrodes. The spent Bayer liquor was pumped from a 12 L reservoir through the anolyte cell inlet and back out to the reservoir. Heat tape wrapped around the bottom of the tank was used to heat the anolyte up to 90 ⁰C. The reservoir and all the piping were insulated to reduce heat loss. A smaller 2 L reservoir was used for the catholyte and was also insulated. Both reservoirs were fitted with condensers at the top that were cooled with tap water to reduce loss of water from the electrolytes. A peristaltic pump was used to deliver water to the catholyte reservoir for those experiments in which the catholyte concentration was held constant.

Unless stated otherwise, for all experiments, flow rate, temperature, inlet pressure, cell voltage, current and charge were monitored.

For experiments 1-12, the anolyte was preheated to 90 ⁰C before transferring to the anolyte reservoir.

Samples of each electrolyte were taken before, during and after each experiment and analysed for total caustic, sodium and aluminium. Sodium gluconate was added to an aliquot of each final anolyte to inhibit precipitation Of AI(OH)₃.

Aluminium content was measured by atomic absorption spectroscopy (wavelength: 309.6 nm, 0.7 nm slit, Al lamp current: 10 mA, acetylene fuel = 1.0, Nitrous oxide oxidant = 2.4).

A four point calibration curve in the range of 5.0 mg/L to 40.0 mg/L was prepared, and anolyte samples containing Al were diluted 1 :2500 and catholyte samples were diluted 1:50.

Sodium content was measured by ion chromatography (Dionex DX320 system with AS40 Autosampler; Analytical Column, 4 mm lonPac CS12A and CG12A Guard Column; Flow Rate: 1.0 mL/min; CSRS Ultra Suppressor, Current: 110 mA; Detector Range: 0.1 μs; Eluent Concentration Gradient: 20 mM MSA (methanesulfonic acid) to 35 mM MSA over 10 minutes; Data Collection Time: 10 minutes).

A four point calibration curve for Na⁺ and K⁺ was prepared over the range of 2.0 mg/L to 100mg/L All anolyte samples were diluted 1 :2500 in deionised water, catholyte samples diluted 1 :5000. All sample quantifications were performed from a linear calibration, and a standard was analysed every 5 to 10 samples and at the end of each sequence.

Total caustic analysis was performed by pH titration, with 0.4997N sulfuric acid (Sigma Aldrich cat. 319570), to a predetermined pH endpoint using a proprietary procedure which takes into account the complete neutralisation of free hydroxide and aluminate-bound hydroxide ([OHT] + [AI(OH)₄ ^']). In this procedure, a 2 ml_ aliquot of liquor anolyte was dispersed in a mixture of 30 ml_ of 400 g/L sodium gluconate solution and 8 ml_ of de-ionised water.

Catholyte samples were titrated using the same procedure.

PART A: Caustic removal from spent liquor by electrolysis using a two compartment cell and a cation permeable membrane

Electrolysis as a function of current density

Experiments 1 , 2 and 3 in Table 1 were performed using potassium hydroxide catholyte in order to monitor the increase in sodium concentration in the catholyte compartment and establish a baseline for the variables studied. Experiment 1 was performed at a current density of 150 mA/cm² using 1M KOH catholyte. The initial AATC ratio was 0.40 and the electrolysis was operated at 90 ⁰C for 6.7 hr to provide a final AfTC ratio of 0.52 and an anolyte TC of 196 g/L Na₂CO₃. The Na⁺ and OH^" current efficiencies were 94.7% and 93.5% respectively. Experiment 2 was performed to determine the effect of a higher current density on cell performance. At 350 mA/cm², the experiment ran for 3.2 hr at 90 ⁰C stopping at an A/TC ratio near that of Experiment 1. The current efficiencies for Na⁺ and OH^" were slightly improved compared to Experiment 1.

Experiment 3 was run under the same conditions as Experiment 2, but to a higher A/TC ratio. At a ratio of 0.62 the current efficiencies were 95 % and 96 % for Na⁺ and OH^" respectively. There was some indication that the current efficiencies decrease slightly at higher A/TC ratios. Although the cell performed well running to a low caustic concentration in the anolyte ( 163 g/L Na₂COs), there was some precipitation of aluminium hydroxide in the sample vials after cooling even after addition of sodium gluconate. The cell was disassembled following Experiment 3 to check for fouling. There was a small amount of build-up of aluminate in the anolyte flow channels and some blistering on the Nafion 324 membrane. The aluminate deposits in the flow frame were believed to be due to the insolubility of alumina at low caustic concentration, and the membrane blisters may have been caused by a number of problems associated with the initial start up and testing of the cell including low flow, high current density, and high A/TC ratio. Membrane blistering was not observed during the remainder of the study.

I Experiment I 1 I I 2 I 3 I

I CD (mA/cm']ι__ j 150 I I 350 350 I

I Membrane I Nafion324 J I Nafion324 | Nafion324 ^l

I Anode | Pt/Nb j I Pt/Nb _J Pt/Nb J

LESDP CQ) I 90 I 90 90 I

|j1mejhr}_____ | 6.7 ! I 3.2 I I

I Charge passed (C) | 360210 [ 389200 ! 548200 I

L[OH] Catholyte Initial (M)__ i _i°J!<9ϋ) I 1-1 <!∞H)___I , 1.1 (KOH) J

I [OH] Catholyte Final (ML___ | 2.7 L_ ²-⁹ !

LJA Anolyte Initial (g/L) J 101 I loo i I 103 I

I TA Anolyte Final (g/y__ [ 101 I 101 __ I L_ ioi

I ^{τs Ano|}yJ§iⁿiM!i9^/LL_ 1 U_ 347 I 345 I I 352 I

I TS Anolyte Final (g/L) | I 314 ^' _, j 324 I 298

LjC Anolyte Injtialjg/L) | 252 L_ 244 I I 244 I

LjTC Anolyte Final (g/Ll____ I I 196 I >177 i I 163 I

L[AI] Catholyte (g/L) [ 0.27 I 0.39 ! L_ °-⁵³

I Na⁺ mass balance I I 98.4 I 103 ! I 94.1

I Al mass balance | I 100 I 101 I I 93.6

I Water transport (mol/mol Na⁺) | L 1.4 ^■ I 1.6 I I 1.5 I

I Average voltage | I 4.33 L— ⁶-⁴⁷ I 6.99 J

I Na⁺ CE 1 L 94.7 [ 98.0 I I 64.6

I OH^" CE ! I 93.5 I 98.8 L_ 96.1 I

Electrolysis as a function of catholyte caustic strength

In order to recycle the caustic catholyte in the Bayer circuit, it would be advantageous if it were highly caustic. Tests were performed at different catholyte (NaOH) concentrations to determine the effect on current efficiency, water transport and hydroxide back migration as shown in Table 2. Experiments 4, 5, 6 and 7 were performed at a catholyte concentration which was close to that at the start by pumping water in to dilute the build up of Na⁺. Results of the four experiments show the water transport across the membrane was between 1.2 and 2.0 mole/mole Na⁺. The Na⁺ current efficiency decreased with increasing catholyte concentration as shown in Figure 3. At caustic concentrations of 7-8 M, the current efficiency was observed to be around 90 % and it was decided to conduct the remainder of the study at catholyte concentrations which were close to this range. Approximately 1.5 % of the charge passed was attributed to Al³⁺ transport and the remainder to OH^" back migration. Note that the performance of this membrane for caustic production is considerably better than would be predicted. It is believed that was due to the high osmotic strength of the electrolytes (particularly the anolyte) which controls the water content of the membrane and limits the back-migration of hydroxide and also accounts for the low water transport rates.

I Al mass balance | I 103 IL 98.J5 I 99.2 L 96.6 ;

I Water transport (mol/mol Na⁺) | I 1-2 1.3 I lβ _

I Average voltage J L ⁵-⁹² (I 6.05 I 6.17 I 6.28 I

I Na⁺ CE _J I 95.3 Il 92.3 I 90.1 I 847__ I

I OH^' CE I 89.2 ( 87.3 L___J-§-§__J

(_A/TCjnitial _J L °-⁴¹⁷ ! 0.420 I 0.400 J J 0.401 !

L ATTC final | I 0.579 _ t| ≤0-⁶l§_—J 0.555 L <0.634 I fable 2. Electrolysis as a function of catholyte caustic strength.

Electrolysis as a function of membrane

Nafion 324 is a reinforced composite of two sulfonate films with different equivalent weight (1100 and 1500) and typically used for producing high concentration NaOH (12 - 20 %). The high equivalent weight layer on the cathode side limits hydroxide back migration. An alternative membrane that could be used in a high strength caustic is a Nafion 400 series membrane, a single layer, lower equivalent weight (1100) sulfonate typically used in the production of strong KOH.

Experiment 8 was performed using Nafion 424 membrane in place of Nafion 324 under the same conditions as Experiment 6 as shown in Table 3. The experiment ran for 4.25 hours passing 536500 coulomb of charge, and transporting 3.6 mole Na⁺ to the catholyte. The average cell voltage with the membrane was lower at 5.2 volts compared to Experiment 6 at 6.2 volts, and the current efficiency was only 65%. The low efficiency was believed to be due to the low equivalent weight membrane being a more open structure allowing a high rate of hydroxide back migration.

Electrolysis as a function of Temperature

It is known to use multiple precipitators operating at different temperatures in Bayer circuits and tests were conducted to determine the effect of temperature in the electrolysis. Experiment 9 was performed at 60 ⁰C with all other conditions the same as Experiment 6 as shown in Table 4. Experiment 9 ran for 3.75 hours using 6.0 L spent liquor anolyte and 25% NaOH catholyte, passing 473400 coulombs of charge. The catholyte concentration was held constant at 8 M NaOH by pumping in water at 1.7 mL/min. The amount of sodium transport to the catholyte was 4.5 moles with efficiency of 88.7 %. The final A/TC ratio was 0.58, and the average cell voltage was 7.2 V, which was expected at the lower temperature. There was no sign of membrane fouling or precipitation of AI(OH)₃ in the cell.

Electrolysis as a function of starting A/TC ratio

To simulate an electrolysis using a liquor having a higher A/TC ratio, aluminium hydroxide was added to the spent liquor so that the ratio was increased The electrolysis was also run at a lower temperature to mimic the temperature in tanks which are situated midway in refinery precipitation banks. For example, in Experiment 11 (see Figure 4), 128 g of AI(OH)₃ was added to 6 L of spent liquor and the mixture was heated to over 90 ⁰C until all of the solid had dissolved. The resulting A/TC was 0.52. The electrolysis was operated at 75 ⁰C with a catholyte caustic concentration of 7.2 M. All other conditions were similar to Experiment 6 as shown in Table 5. The initial AI₂O₃ concentration was 112 g/L, and the total caustic was 214 g/L Na₂CO₃. The electrolysis ran for 2.2 hr, passing 273500 coulombs of charge, increasing the A/TC ratio to 0.665. The amount of sodium transported to the catholyte was 2.6 mol and the resulting final catholyte concentration was 7.6 M NaOH, which was held constant by adding water. The Na⁺ current efficiency was 91 %. The result demonstrates that it should be able to supersaturate liquor of any A/TC, providing that it is done under conditions which inhibit the spontaneous precipitation of alumina.

Electrolysis as a function of electrode material

Various materials were tested for use as suitable anodes as shown in Table 6. Stainless steel was used as a cathode for the electrolysis experiments performed and showed no signs of degradation throughout the study. Platinum coated niobium which is a stable anode was used when investigating electrolysis conditions. Several materials were tested in a small undivided glass cell using spent liquor as an electrolyte. Nickel, stainless steel 316 and DSA O₂ (Eltech EC600) were tested at 250 mA/cm² at 80 ⁰C over time. The DSA (dimensionally stable anode) has an IrO₂ coating that had been stripped off after 2 hr in the glass cell, but both Ni and stainless steel appeared to be in good condition. The glass cell containing the SS 316 anode continued electrolysing for 25 hr in liquor with no apparent damage.

A stainless steel anode was installed in the electrolysis cell for further testing. Experiments 10, 11 and 12 were performed with this anode with no visible signs of deterioration. Cell performance for each experiment was good with efficiencies ranging from 87 to 91 %.

Further testing on the anolyte was conducted to determine if any corrosion not apparent by visual inspection was occurring. Samples from each of Experiments 10, 11 and 12 and also samples from a Pt/Nb anode experiment (Experiment 8) were analysed for iron by atomic absorption. Figure 5 shows the Fe concentration for Experiments 10, 11 and 12 increased, believed to be a result of corrosion at the anode surface. Experiment 12 was performed at a 150 mA/cm² to determine if the corrosion was a function of current density. Based on Fe analysis, the corrosion rate was higher at the lower current density, but lower when run at 75 ⁰C (Experiment 11) as opposed to 90 ⁰C.

Additional experimentation was conducted using a small divided glass cell to screen other materials for use as a suitable anode. In Figure 6, Nickel, Monel and Hastelloy C were tested as anodes in a divided glass cell using 50 ml_ of spent liquor anolyte and 25% NaOH catholyte. Each sample was electrolysed at 350 mA/cm² for 1 to 2 hours at 90 ⁰C. Each of the anolytes was analysed for Ni by atomic absorption. Hastelloy C is a heat resisting alloy used in chemical processing and pulp and paper production and is known for its high corrosion resistance at high temperatures. However, under the experimental conditions, corrosion was apparent both visually and quantitatively. Both nickel and Monel anodes also corroded slightly under the same conditions. Nickel was detected at 15 mg/L for the Ni anode and 13 mg/L using the Monel anode.

In Experiment 13, the SS 316 anode was replaced with a platinised titanium anode and run under standard conditions (as in Experiment 6 and presented in Table 7). Running at 35 A, 5.5 moles charge was passed removing 4.89 moles OH^" based on the increase in catholyte concentration, resulting in 89 % efficiency. The cell was disassembled following the experiment to inspect the anode. The anode appeared to be worn (i.e. some Pt loss), however the extent of corrosion was not determined.

Further work which focussed on testing amorphous carbon electrodes also showed these to corrode significantly with time. On the other hand, platinised niobium electrodes, when electrolysed at 350 mA/cm² and 70 ^°C, showed no signs of corrosion after more than 500 hr of testing.

Discussion of Results

The experiments demonstrate that the removal of caustic from a Bayer spent liquor by electrolysis, using only a cation permeable membrane can be achieved with high (90 %) efficiency while producing up to 25 % NaOH which can be recycled back within the Bayer circuit. The 10 % inefficiency is believed to be due to back migration of hydroxide and aluminium transport to the catholyte, with 1.5 to 2 % of the charge being due Al³⁺ transport. 90% efficiency was obtained under the following conditions:

current density 350 mA/cm²;

temperature 90 ⁰C; and

catholyte concentration 8M NaOH.

Under these conditions 70 g/L TC as Na₂CO₃ can be removed from the liquor increasing the A/TC ratio to 0.65 without short term fouling of the membrane, or AI(OH)₃ precipitation problems within the cell. The average cell voltage was 6 volts and varies with electrode and current density. It is noted that the inter- electrode gap in the experimental cell is considerably larger than would be the case in a commercial electrolysis cell and it is expected that the cell voltage would be 1 -2 volts less.

EXAMPLE 1

The following example demonstrates that spent liquor can be supersaturated by the removal of soda using membrane electrolysis, and seeded to produce significant amounts of alumina compared to a control sample of spent liquor that has not undergone prior electrolysis.

A sample of liquor ex-precipitation (LXP) from the Applicant's Kwinana refinery, having a TC of 241 (expressed as grams per litre of sodium carbonate) and an A/TC of 0.48 was split into two portions. One of the portions was subjected to electrolysis at 90 ⁰C, using the Nafion 324 membrane, to reduce the TC to 167 g/L and increase the A/TC to 0.68. Both the LXP and the electrolysed liquor (100 mL of each) were transferred to separate polypropylene bottles and gibbsite seed (10 g) added to each solution. The bottles were sealed and placed in a rotating water bath at 70 ⁰C for 24 hr after which time the solids were collected by filtration, dried at 105 ⁰C, and weighed. Both filtrates were analysed for alumina and TC content. The results for both experiments are given in Table 8 where Experiment 14 refers to the electrolysed spent liquor and Experiment 15 refers to the untreated LXP (control experiment). Clearly, the electrolysed spent liquor has produced a significant yield increase compared to the control, which equates to an extra 24.6 g/L additional yield of hydrate.

The example demonstrates that LXP can be re-supersaturated by the removal of soda using membrane electrolysis, and that substantial yields of alumina can be obtained by additional seeding of the liquor.

I Experiment . M I I 15 ,

JJTC Anolyte Initial (g/L) | 167 i I 241 I

I Seed added (g) I 10

|_TC Anolyte Final Filtrate jg/L) _J 209 I I 249 i

LA/TCJnitial | 0.683 I L 0.480 I

LA/TCJinal I 0.462 I 0.458 I

I AI₂O₃ initial (as hydrate) (g/L) j HUH^{4 as A|}2θ₃) I L ¹⁷⁷ (U£^{as A|}2°3) J

I AI₂O3 hydrate precipitated (g)* j 2.647 _J L 0.184

Table 8. Alumina precipitation results.

* solid recovered, corrected for seed hydrate added

PART B: Caustic removal from spent liquor by electrolysis using a two compartment cell, gas diffusion electrodes and a cation permeable membrane.

Reductions in cell voltage and hence power costs may be achieved by utilising gas diffusion electrodes instead of conventional flat plate electrodes.

At hydrogen depolarised anodes, the oxidation of hydrogen occurs according to:

H₂ ► 2H⁺ + 2e^" E⁰ = 0.0 V vs SHE (at pH 0)

This reaction occurs at significantly lower potential than the alternative anodic oxygen producing reaction that occurs at flat plate electrodes according to:

2H₂O ► O₂ + 4H⁺ + 4e^' E⁰ = 1.23 V vs SHE (at pH 0)

Similarly, at oxygen depolarised cathodes, the reduction of oxygen occurs according to: O₂ + 2H₂O + 4e ^ 4OH- E⁰ = 0.401 V vs SHE (at pH 14)

Compared to the alternative cathodic hydrogen producing reaction:

H₂O ► H₂ + 2OK E⁰ = 0.8277 V vs SHE (at pH 14)

In addition, overpotentials required for gas evolving reactions are relatively high, further adding to the energy savings that can be achieved through the use of gas diffusion electrodes.

Gas diffusion electrodes

The experiments were carried out in a Microcell (Electrocell AB, Sweden) with an active surface area of 10 cm². Cell configurations tested were:

1. Gas diffusion electrode as an anode as shown in Figure 7;

2. Gas diffusion electrode as a cathode as shown in Figure 8; and

3. A conventional set up employing flat plate electrodes as shown in Figure 9.

A Nafion 350 membrane, employing a bi-layer structure with higher equivalent weight polymer facing the cathode to minimise back migration of hydroxide ions was used in all of the configurations. The gas diffusion electrode was an ELAT^R LT 140E-W SI (E-TEK₁ New Jersey, USA). The ELAT^R electrode has a Nafion coating facing the solution to minimize solution breakthrough. Spent liquor from the applicant's refinery at Kwinana, was used as anolyte and 5 % NaOH was used as catholyte. Voltage versus current density curves were generated at temperatures of 40, 50 and 60 ⁰C for each configuration and the results are presented in Figures 10-12

In all cases, it can be seen that the cell voltage of the solutions increase as the solution temperatures decrease due to decreasing conductivity. Importantly, it is evidenced from Figure 13 that when the gas diffusion electrode was used, either as the anode or the cathode, a voltage reduction of approximately 1.1 V was achieved with respect to when no gas diffusion electrode was used at all. This lower voltage equates to a substantial reduction in energy consumption.

A further electrolysis experiment was performed on Kwinana spent liquor at a constant temperature of 60 ⁰C and a current density of 167 mA/ cm². The results are summarised in Table 9.

Table 9.

This experiment demonstrates that it is possible to recover 8% caustic in the catholyte compartment at a current efficiency of 85 % whilst maintaining very low levels of aluminium migration across the membrane (only 89 mg/ L of Al was detected in the catholyte). Furthermore, the A/TC ratio of the spent liquor increased from 0.35 to 0.71 during the run which signifies a substantial amount of supersaturation with respect to alumina solubility. This was achieved at a cell voltage ranging between 3-3.4 V which is much lower than that obtained without the use of a gas diffusion electrode (i.e. with conventional flat plate electrodes).

PART C: Caustic removal from spent liquor by electrodialysis using a multi-compartment cell containing cation permeable and bipolar membranes.

The cell used for the bipolar membrane electrodialysis was a Eurodia

Electrodialysis stack (EUR2B-9) with 0.2 m² effective electrode area each side. The cell was built with a platinised titanium anode and a stainless steel cathode along with 9 pairs of Fumatech FKB cation membranes 62 and Neosepta bipolar BP-1 membranes 64, and Nafion 115 membrane at both ends 66, see Figure 14. The feed (LXP liquor) 24 and concentrate (10 % sodium hydroxide) 33 were pumped through separate cell compartments at approximately 3 L/min (0.3L/min/comp.). An electrode rinse (NaOH) 68 was provided adjacent the electrodes.

A separate electrode rinse consisting of 0.2M sodium hydroxide was pumped to both the anode and cathode and combined at the cell outlet in a tank where the two electrolytes were degassed. The cell was operated at constant voltage of 26 volts (2.1 V/cell + 3 V/electrode).

A Teflon coated immersion heater was used to heat the spent liquor anolyte feed to the desired temperature (either 40 or 60 ⁰C).

Experiment 17, Table 10, was operated at a constant temperature of 40 ⁰C, constant cell voltage of 26 V and 10% sodium hydroxide concentrate in the catholyte which was held constant with the addition of water at 15 mL/min. The feed flow rate was 3L/min with a pressure of 4.4 psi measured at the cell inlet, and 2.9 L/min for the concentrate flow with a back pressure of 4.4 psi. The electrodialysis stack was operated at equal inlet pressures to prevent cross flow leaking between flow compartments.

The experiment ran for 6.5 hr with several samples taken during the run for analysis. The spent liquor feed total caustic concentration decreased from 237 to 142 g/L as Na₂CO₃ with a current efficiency of 74%. The average current density was 56mA/cm², and the A/TC ratio for the LXP liquor increased from 0.322 to 0.516, see Figure 15.

Experiment 18 was operated under the same conditions as Experiment 17, except at a higher temperature of 60 ⁰C. The experiment ran at a constant 26 volts for 7 hours decreasing the total caustic in the spent liquor from 248 g/L to 129 g/L as Na₂CO₃ with a current efficiency of 83%. Water was added to the catholyte to hold the concentration constant which started with 4 L of 3.4M (12.3%) NaOH and finished with 12.9 L of 2.8M (10.3%) NaOH. The AI₂O₃ spent liquor concentration decreased by 7 % from 76 g/L to 71 g/L caused mostly by dilution of the spent liquor due to water transport across the FKB membrane, with a small amount of AI₂O₃ transport to the caustic, see Figure 16. The spent liquor feed volume increased from 10 L to 10.6 L, but the total mass of aluminum dropped very little form 14.9 moles to 14.7 moles. The A/TC ratio increased from 0.306 to 0.551 , and the average current density was 60mA/cm².

PART D: Caustic removal from spent liquor by electrolysis using a multicompartment cell containing cation permeable membrane and anion permeable membranes.

Electrodialysis involves the transportation of ions through membranes under the influence of an electric field. Importantly, if alternating cation and anion permeable membranes are used in a multi-compartment cell, ions can be transported into adjacent compartments without the splitting of water to maintain charge balance. For example, when Kwinana liquor is pumped through the feed compartment of an electrodialysis stack and an electrical potential is applied between the anode and the cathode, the positively charged cations, such as sodium, migrate through the cation permeable membrane toward the cathode and the negatively charged anions migrate through the anion permeable membrane towards the anode.

The cation permeable membrane rejects the passage of anions (OH^", Cl^", AI(OH)₄ ^" , SO₄ ^") and the anion permeable membrane rejects the passage of cations.

The overall result is a decrease in the NaOH concentration in the liquor, or feed stream, and an increase in the NaOH concentration in the concentrate stream.

This type of arrangement can result in energy savings compared to arrangements requiring the splitting of water to maintain charge balance. The configuration used in the following experiments is shown in Figure 17.

The entire electrolysis set-up was contained within a fume hood, for proper venting, and samples of liquor were withdrawn regularly and analysed for total caustic, aluminium, suphate and chloride.

The electrodialysis runs were performed in a FuMaTech FT-ED100-4-10 (Fuma- Tech GmbH, Germany). The stack consisted of a DSA-O2 anode, platinised titanium cathode and a combination of Neosepta AHA anion 70 and Fumatech FKL cation 72 exchange membranes. The Neosepta AHA membranes have a high mechanical strength and are base stable. The FKL membranes were chosen for their low hydroxide leakage properties. There were 10 ED membrane pairs each with an operating surface area of 0.01 m². The feed compartment consisted of a 2.5 L polypropylene reservoir with a 300 W PTFE coated immersion heater, a 5 micron polypropylene filter and an Iwaki WMD-30LFX centrifugal circulating pump. The inlet pressure and solution temperature (maintained at 60 ⁰C) were monitored during the run.

The concentrate or base loop consisted of a 2 gallon polypropylene reservoir and a Iwaki WMD-30LFX centrifugal circulating pump. The inlet pressure and flow rate was monitored during the run. Depending on the run, 5 % or 9 % caustic was used as the starting concentrate. Deionised water was metered in slowly during the run to maintain a constant concentration of base.

The electrode rinse loop consisted of a 2 L PTFE reservoir and an Iwaki WMD- 30LFX centrifugal circulating pump. The electrode rinse solution (0.05 M Na₂SO₄) was split into two streams before entering the anode and cathode compartments. The solutions exiting the compartments were recombined in the main reservoir. It was anticipated that this configuration would maintain pH neutrality in the rinse solution. The electrode reactions are shown below.

Cathode: H₂O + e ^{" ►} V₂H₂ + OH^'

Anode: H₂O * V₂O₂ + 2H⁺ +2e^"

The electrode rinse solution was maintained at a pH of 2.5 - 4 by the addition of concentrated sulfuric acid. Power was supplied to the stack by a GW Model GPR-1810HD DC power supply. The cell voltage was monitored and recorded during the run and several samples of each stream were taken for later analysis.

After each experiment starting with Experiment 23, the feed side of the cell was washed with 10 % NaOH for approximately one hour at temperatures up to 55 ⁰C.

Electrodialysϊs at a constant catholyte 5 % ^w/_w caustic concentration and different current densities

Results pertaining to the electrodialysis of spent Bayer liquor, from the applicant's Kwinana refinery are summarised in Table 11. In this series of experiments (numbers 19-21), the caustic concentration in the catholyte compartment was maintained at approximately 5 % w/w. It is evident that current efficiencies ranging between 71.4 to 74.8 % were obtained at A/TC ratios of up to 0.79. The main inefficiency is believed to be caused by back-migration of hydroxide species across the cation exchange membrane, accounting for approximately 12-17 % of the charge. The back migration was estimated based on the amount of acid neθded to maintain a constant pH in the electrode rinse compartment. The transport of other anionic species present in the liquor accounted for the remainder of the charge.

Figure 18 shows a typical concentration profile of spent Bayer liquor during the electrodialysis experiments. Notably, nearly all of the chloride in the liquor is removed during the early stages of the run, which accounts for ~ 6 - 8 % of the total charge. Negligible amounts of sulphate are also removed (< 1 % of total charge) and the typical cell voltages attained for the removal of caustic to give a catholyte concentration of 5 % w/w NaOH, for two different current densities, are shown in Figure 19.

The profile outlined in Figure 19 indicates that, irrespective of current density, the voltage passes through a minimum as the caustic content of the treated liquor continues to fall. A voltage of 0.5 V / membrane pair at 50 mA/ cm², and 1 V / membrane pair at 100 mA/ cm², (assuming 3 volts for the end electrode reactions) was measured at the minimum. Analysis of samples at the minimum voltage indicated an A/TC ratio of 0.46 - 0.48. Water transport values were measured at between 1 to 1.7 moles water per mole of charge passed.

Electrodialysis of spent Bayer liquor at a constant 9% w/w catholyte caustic concentration

Results for the electrolysis of Kwinana spent liquor at a constant catholyte caustic concentration of 9 % w/w, 50 mA/ cm² current density and different final A/TC ratios are presented in Table 12 (Experiments 22 - 23). There is a notable reduction in current efficiency (down to between 62 -63 %) compared to the results obtained for a catholyte caustic concentration of 5 % w/w (71 - 75 %) which is due to a higher level of hydroxide back migration. There was also a decrease in cell voltage (down to 0.4 V / membrane pair as opposed to 0.5 V / membrane pair) due to a higher conductivity of the catholyte. Both the current efficiency and voltage did not change significantly during the removal of caustic from the anolyte up to an A/TC ratio of 0.79.

Table 12. Electrodiaiysis of spent Bayer liquor at a constant catholyte 9 % w/w caustic concentration The results for experiments performed at two different current densities, 50 mA/ cm² (Experiment 23) and 100 mA/ cm² (Experiment 25), at a constant catholyte concentration of 9 % w/w caustic, are presented in Table 13.

concentration and different current densities

The current efficiencies for both experiments are very similar, at close to 63 %. However, there does appear to be a higher level of hydroxide back migration into the anolyte compartment at the lower current density. Also, there is a notable difference in the amount of water transport into the catholyte compartment at the higher current density, which is significantly higher at 2.2 moles per mole of charge compared to 1.2 moles per mole of charge. At 100mA/ cm² current density, the overall cell voltage was lower (0.8 V/ membrane pair, Experiment 25) when the catholyte contained 9 % w/w caustic as opposed to 5 % w/w caustic (1 V/ membrane pair for Experiment 21 ).

The typical cell voltages for the removal of caustic soda at 9% w/w in the catholyte are shown in Figure 20 Analysis of samples at the minimum voltage indicated an A/TC ratio of 0.49 to 0.51. In several experiments, the cell voltage spiked substantially after going through a minimum. This may have been due to the precipitation of solids within the cell as the A/ TC ratio of the anolyte increases beyond a critical point during electrolysis. Overall, electrodialysis of spent Bayer liquor using cells containing both cation and anion permeable membranes has been successfully demonstrated, as a means of supersaturating Bayer liquor with respect to alumina solubility, at current densities up to 100 mA/ cm² and cell voltages of between 0.5-1 V per membrane pair. Caustic soda strengths of up to 9 % w/w were recovered at current efficiencies of between 62 -63 %. Higher efficiencies can be achieved by lowering the strength of caustic recovered. The loss in efficiency is somewhat offset by the increased conductivity of the higher strength caustic produced and voltage savings of between 0.1-0.2 volts per membrane pair can be achieved (depending on the current density) by increasing the strength of the caustic product from 5 to 9 % w/w.

Importantly, it has been shown that spent Bayer liquor having a low A/ TC ratio of, for example 0.34, can be supersaturated to A/TC ratios above 0.7. Therefore, the treated liquor could be seeded to precipitate more alumina product.

The major advantage of employing electrodialysis for the removal of caustic from Bayer liquor, as opposed to electrolysis, is that the former is considerably less energy intensive. An example of the comparative power cost, expressed as per tonne of alumina produced, associated with caustic extraction and re- supersaturation of spent liquor using both techniques is shown in Table 14. The data is based on obtained experimental data.

electrolysis versus electrodialysis of Bayer liquor Modifications and variations such as would be apparent to a skilled addressee are deemed to be within the scope of the present invention.

Claims

The Claims Defining the Invention are as Follows:

1. A method for controlling the precipitation of alumina from a Bayer process solution, the method comprising the steps of:

applying a potential between a first region comprising a Bayer. process liquor and a second region comprising a caustic solution, wherein an ion permeable membrane is provided between the first region and the second region; and

wherein the Bayer process liquor is not directed to the second region.

2. A method for controlling the precipitation of alumina according to claim 1 , wherein there are provided more than two regions and more than one ion permeable membrane.

3. A method for controlling the precipitation of alumina according to claim 1 or claim 2, wherein the caustic solution has a maximum alumina concentration of 20 g(_^'1 as AI₂O₃.

4. A method for controlling the precipitation of alumina according to claim 2, wherein there is provided at least one anion permeable membrane and at least one cation permeable membrane.

5. A method for controlling the precipitation of alumina according to any one of claims 2 to 4, wherein the plurality of ion permeable membranes comprise an electrodialysis unit.

6. A method for controlling the precipitation of alumina according to any one of the preceding claims, wherein there is provided a bipolar membrane.

7. A method for controlling the precipitation of alumina according to claim 2, wherein there is provided at least one bipolar membrane and at least one cation permeable membrane.

8. A method for controlling the precipitation of alumina according to any claim 6 or claim 7, wherein the plurality of permeable membranes comprise an electrodialysis unit.

9. A method for controlling the precipitation of alumina according to any one of the preceding claims, wherein the method comprises the further step of:

precipitation of alumina in the Bayer process liquor.

10. A method for controlling the precipitation of alumina according to any one of the preceding claims, wherein the method comprises the further step of:

seeding the Bayer process liquor with alumina.

11.A method for controlling the precipitation of alumina according to claim 10, wherein the step of:

seeding the Bayer process liquor with alumina,

is conducted prior to the step of:

applying a potential between a first region comprising a Bayer process liquor and a second region comprising a caustic solution, wherein an ion permeable membrane is provided between the first region and the second region

12. A method for controlling the precipitation of alumina according to claim 10, wherein the step of:

seeding the Bayer process liquor with alumina ,

is conducted after the step of:

applying a potential between a first region comprising a Bayer process liquor and a second region comprising a caustic solution, wherein an ion permeable membrane is provided between the first region and the second region.

13. A method for controlling the precipitation of alumina according to any one of the preceding claims, wherein the step of:

is conducted at a temperature which is below that used to digest the bauxite.

14. A method for controlling the precipitation of alumina according to any one of claims 9 to 13, wherein the step of:

precipitation of alumina in the Bayer process liquor;

is conducted at a temperature up to the boiling point of the liquor at that pressure.

15. A method for controlling the precipitation of alumina according to any one of claims 4 to 14, wherein the cation permeable membrane comprises perfluorinated polymers such as a sulfonated tetrafluorethylene copolymer, carboxylate polymer, polystyrene based polymer, divinylbenzene polymer, or sodium conducting ceramics such as beta-alumina or combinations thereof.

16. A method for controlling the precipitation of alumina according to any one of claims 4 to 14, wherein the anion permeable membrane is a Neosepta AHA membrane or a Fumatech FAP membrane.

17. A method for controlling the precipitation of alumina according to any one of the preceding claims, wherein the anode comprises platinum coated niobium, platinum coated titanium or Monel.

18. A method for controlling the precipitation of alumina according to any one of the preceding claims, wherein the cathode comprises stainless steel or a gas diffusion electrode (oxygen depolarized cathode).

19. A method for controlling the precipitation of alumina according to any one of the preceding claims, wherein the caustic concentration is not greater than about 8M NaOH or 25% NaOH.

20. A method for controlling the precipitation of alumina according to any one of the preceding claims, wherein the caustic solution is sourced form within the Bayer circuit.

21. A method for controlling the precipitation of alumina according to any one of the preceding claims, where the Bayer process includes the steps:

digestion of bauxite with caustic solution;

liquid-solid separation to provide a residue and a Bayer liquor; and

precipitation of alumina from the Bayer liquor;

the steps of:

are conducted prior to the step of: precipitation of alumina from the Bayer liquor.

22.A method for controlling the precipitation of alumina according to any one of claims 1 to 20, where the Bayer process includes the steps:

digestion of bauxite with caustic solution;

liquid-solid separation to provide a residue and a Bayer liquor; and

precipitation of alumina from the Bayer liquor; and

the steps of:

are conducted after the step of:

precipitation of alumina from the Bayer liquor.

23.A method for controlling the precipitation of alumina according to any one of the preceding claims, wherein the method is performed as a batch process or a continuous process.

24.A method for controlling the precipitation of alumina substantially as hereinbefore described with reference to the Examples.

25. A method for controlling the precipitation of alumina substantially as hereinbefore described with reference to the Figures.