THIS INVENTION relates to the treatment of water which contains dissolved calcium, sulphate and hydroxide ions. It relates in particular to a method of treating such water.

Acid mine drainage and other industrial effluents containing dissolved calcium sulphate ions can be treated to remove the dissolved calcium and sulphate ions by adding barium and carbonate ions in the form of barium carbonate slurry and co-precipitating barium sulphate and calcium carbonate as solid particles. This is usually effected in a single reactor such as a mixed tank reactor or a column reactor, so that the solid barium sulphate and calcium carbonate particles are mixed and are hence withdrawn as a mixed barium sulphate and calcium carbonate stream. Complex and costly downstream processing of the stream or sludge is required, particularly when it is desired to reduce the barium sulphate to barium sulfide, since the presence of calcium carbonate inhibits and/or reduces the efficiency of the conversion of barium sulphate to barium sulfide. Additionally, the reactors have slow reaction kinetics, making the process economically less attractive.

It is hence an object of this invention to provide a method of treating such water, whereby these drawbacks are at least reduced.

Thus, according to the invention, there is provided a method of treating water containing dissolved calcium, sulphate and hydroxide ions, which method includes allowing a reaction medium comprising the water which contains dissolved calcium, sulphate and hydroxide ions, dissolved barium ions and dissolved carbonate ions, to react with carbon dioxide in a reaction zone, to form, as precipitates, solid calcium carbonate particles and solid barium sulphate particles, with the calcium carbonate particles having a higher settling velocity than the barium sulphate particles;

maintaining, in the reaction zone, a fluidized bed of the calcium carbonate particles, with the barium sulphate particles passing upwardly from the reaction zone; and

allowing calcium carbonate particles to pass downwardly from the reaction zone.

The method may thus include introducing the barium ions into the reaction zone. This may be effected by introducing, e.g. by adding, a solid barium compound to the reaction zone and allowing at least some of the particulate barium compound to dissolve in the water of the reaction medium, thereby releasing barium ions into the water, and hence into the reaction medium. In particular, the solid barium compound may be particulate barium carbonate which partly dissolves in the water so that, in addition to soluble barium ions being released into the water, soluble carbonate ions are also released into the water, and hence into the reaction medium; unreacted barium carbonate particles then also pass upwardly from the reaction zone together with the barium sulphate particles. The method may include introducing, eg feeding, the water containing the dissolved calcium, sulphate and hydroxide ions, into the reaction zone.

The reaction zone may be a first reaction zone, with a second reaction zone also being provided. The process may then include allowing the barium sulphate particles and unreacted barium carbonate particles to pass from the first reaction zone into the second reaction zone.

The method may include withdrawing, from the second reaction zone, some of the water, most of the unreacted residual barium carbonate particles, and some of the barium sulphate particles, and allowing the barium sulphate particle containing water to pass into a third reaction zone. The method may then also include withdrawing barium sulphate particles from the third reaction zone.

The method may further include withdrawing calcium carbonate particles from the first reaction zone.

The first reaction zone may be provided by an upflow fluidized bed reactor which may typically have an elongated upright configuration. The second reaction zone may also be of an elongated upright configuration and having a lower upflow velocity than the first reaction zone, while the third reaction zone will have a lower upflow velocity than the second reaction zone. Typically, the first and second reaction zones may be provided in a single elongated upright reaction vessel, with the first reaction zone provided at the lower end of the reactor vessel and with the second reaction zone located above the first reaction zone, ie at the upper end of the reactor vessel. The third reaction zone may be provided by a separate vessel. Typically, the calcium carbonate particles may be in the form of more-or-less spherical pellets having particle sizes or diameters in the range of 0.1 mm and 1 mm, typically in the region of 0.4mm to 0.5mm. The barium sulphate particles will normally be smaller than the calcium carbonate particles or pellets, typically about 20μm.

The upward liquid velocity in the first reaction zone or fluidized bed region of the reactor will naturally be sufficiently high to achieve fluidization of the calcium carbonate pellets, but naturally not so high as to achieve significant carryover or migration of the calcium carbonate pellets into the second reaction zone together with the barium sulphate and unreacted barium carbonate particles. Thus, when the calcium carbonate pellets are around 0.4mm to 0.5mm in size as hereinbefore described, the upward liquid velocity in the first reaction zone of the reactor can be in the range of 15 meters per hour to 50 meters per hour, typically about 28 meters per hour ('m/h'), while the upward liquid velocity in the second reaction zone of the reactor, upstream of the point of withdrawal of the water containing the unreacted barium carbonate particles and some of the barium sulphate particles, can be in the range of 5m/h to 28m/h, typically about 12m/h, and downstream of this withdrawal point, in the range of 0.5 m/h to 12 m/h, typically about 5m/h. The third reaction zone may have an upflow velocity in the range of 0.01 m/h and 12 m/h, typically less than 3 m/h.

The method may include initially seeding the first reaction zone with a seeding medium, to facilitate calcium carbonate particle formation. The seeding medium may comprise calcium carbonate particles, such as calcium carbonate pellets; fine silica sand; or any other fine particulate material whose size usually does not exceed 0.75mm, and on which calcium carbonate can precipitate. Typically, the seeding medium particles are 0.1 mm to 0.75mm in size.

It is envisaged that the invention will have particular application to the treatment of water which is derived from, or is, acid mine drainage or other effluents containing dissolved calcium sulphate (CaSO ₄ ) and dissolved calcium hydroxide (Ca(OH) ₂ ) as the dissolved calcium ions, and containing no barium ions. The method will then include introducing into the reactor, together with the water, a source of barium ions, particularly a slurry of barium carbonate (BaCOa) particles in water. Typically, the water to be treated is then an intermediate product in a process to treat acid mine drainage (AMD) or effluent with lime to precipitate metals such as magnesium and which raises the water pH, typically to around 11. Thus, addition of the carbon dioxide to the reactor in accordance with the invention, lowers the water pH, typically to around 9. The following reactions accordingly take place in the reactor:

BaCO ₃ + Ca ²⁺ + SO ₄ ^2" <-► CaCO ₃ (S) + BaSO ₄ (S) (1 )

Ca ²⁺ + OH ^" + CO ₂ <→ CaCO ₃ (S) + H ⁺ (2)

In accordance with the invention, CaCO ₃ pellets are thus separated from finer BaSO ₄ particles, with the CaCO ₃ pellets being withdrawn from the first reaction zone in the form of CaCCb sludge and the BaSO ₄ particles being withdrawn from the third reaction zone in the form of a BaSO ₄ sludge.

The main principle behind the separation in accordance with the invention is the degree of supersaturaton of the two precipitates. The solubility of calcium carbonate in water is significantly higher than that of barium sulphate, and therefore it is possible to grow calcium carbonate crystals while it is unlikely that significant barium sulphate crystals growth will occur. The result is that it is possible to produce calcium carbonate crystals of greater size, ie as pellets, than the barium sulphate precipitate, which allows for separation of the two precipitates based on the difference in their settling velocities.

The method may include reducing the calcium ion concentration in the reaction medium before and/or during the reaction thereof with the carbon dioxide, typically to between 200mg Call and 400mg CaIL This may be achieved by recycling water that is low in calcium ions. The reason for this is to reduce the calcium carbonate precipitation potential and therefore the degree of supersaturation of the feed to the column so that nucleation of the calcium carbonate does not take place, but rather crystal growth on the existing solid seeding material present in the column. Barium carbonate is dosed in stoichiometric proportions to the calcium sulphate, and therefore by reducing the calcium concentration the barium carbonate concentration is also reduced. More particularly, the method may include recycling, to the first reaction zone, the barium carbonate and barium sulphate containing water withdrawn from the second reaction zone. This recycled water will contain little or no CaCθ3, but will contain BaSO ₄ , and unreacted BaCθ3. The invention will now be described in more detail with reference to the accompanying drawings and the specific examples set out hereinafter. In the drawings,

FIGURE 1 shows, in simplified flow diagram form, an installation in which a method of treating dissolved calcium and sulphate ion containing water, in accordance with the invention, can be carried out;

FIGURE 2 shows the sulphate concentration and pH of the feed water of Example 2;

FIGURE 3 shows the total carbonate concentration and pH of the carbon dioxide water of Example 2;

FIGURE 4 shows the column effluent sulphate concentration and pH for Example 2;

FIGURE 5 shows, for Example 2, the feed water and effluent pH, demonstrating the effects of high feed water pH on the effluent while a constant acid load was applied to the system;

FIGURE 6 shows a simplified process flow diagram of an installation for carrying out the method of the invention according to Example 3;

FIGURE 7 shows, schematically, a longitudinal sectional view of a reactor used in Example 3;

FIGURE 8 shows, for Example 3, sulphate and calcium concentrations in the pH 11.2 water for period 1 (sulphate=2394±187 (41 points); calcium=972±281 (30 points));

FIGURE 9 shows, for Example 3, sulphate and calcium concentrations in the pH 1 1.2 water for period 2 (sulphate=2418±374 (191 points); calcium=926±356 (236 points));

FIGURE 10 shows, for Example 3, sulphate and calcium concentrations in the pellet reactor effluent for period 1 ;

FIGURE 1 1 shows, for Example 3, sulphate, pH and conductivity of the reactor effluent, showing the relationship between sulphate and conductivity, but no dependence on pH;

FIGURE 12 shows, for Example 3, sulphate and calcium concentrations in the pellet reactor effluent;

FIGURE 13 shows, for Example 3, sulphate concentrations in the effluent and barium concentration in the barium feed tank; and

FIGURE 14 shows, for Example 3, calcium and total carbonate species concentrations in the effluent. Referring to Figure 1 , reference numeral 10 generally indicates an installation in which a method of treating calcium, sulphate and hydroxide ion containing water, in accordance with the invention, can be carried out.

The installation 10 includes an upright reactor 12. The reactor 12 is circular in cross-section and comprises a lower section 14 of relatively small cross- sectional area (and which constitutes a first reaction zone), an outwardly flaring intermediate or clarifying section 16 at the upper end of the lower section 14, and an upper section 18 (and which constitutes a second reaction zone), which is thus of larger cross-sectional area than the lower section 14, above the intermediate section 16.

A feed water line 20 leads into the reactor 12 at its lower end, as do a CO ₂ water addition line 22 and a BaCU ₃ slurry addition line 24. A recycle water line leads from the second reaction zone 18 of the reactor 12, to the line 24. An effluent withdrawal line 28 leads from the upper reactor section 18 to a separator 30 (which constitutes a third reaction zone), with a BaSO ₄ withdrawal line 32 leading from the bottom of the separator 30 and an effluent withdrawal line 34 leading from the upper end of the separator 30. A CaCU3 withdrawal line 36 leads from the lower end of the reactor 12.

The installation 10 typically forms part of an acid mine drainage (AMD) desalination process which includes, upstream of the installation 10, a primary lime treatment stage in which lime is added to the AMD to raise its pH to about 9, with a metal sludge and gypsum precipitating and being removed. The water from the primary lime treatment stage, which is thus at a pH 9, passes to a secondary lime treatment stage where the pH of the water is further raised to about 11 , and in which magnesium hydroxide and gypsum precipitate and are removed. The resultant treated water, which is typically at a pH of 11.2, then passes along the feed water line 20 into the reactor 12.

On start-up of the reactor 12, it is seeded with calcium carbonate pellets having a size or diameter in the range of 0.4mm to 0.5mm. These pellets thus initially are present as a fixed bed at the bottom of the reactor 12 immediately prior to water entering the reactor. Simultaneously with feeding the feed water to be treated into the reactor 12, CO2 water is fed into the reactor 12 along the line 22, as is a BaCU ₃ slurry comprising BaCU ₃ particles suspended in water, along the line 24. The various streams are introduced into the reactor 12 at a sufficient rate so that the upward liquid velocity in the lower section or first reaction zone 14 of the reactor 12 is about 28m/h. This velocity is sufficient to fluidize the bed of CaCθ3 pellets, and to maintain them in fluidization. The reactions in accordance with Reactions (1 ) and (2) hereinbefore set out, take place in the lower reactor section 14, with CaCθ3 crystals growing on the CaCU3 pellets as CaCU3 precipitates out of the reaction medium comprising the feed water, the BaCθ3 slurry and the CO2 water, and with new pellets forming, as do barium sulphate nucleated particles which are smaller than the CaCCb pellets and have a lower settling velocity than the CaCU3 pellets. At the upward liquid velocity of 28m/h, effectively all CaCU3 pellets remain in the lower section of the reactor 12, ie there is little or no carryover of precipitated CaCCb into the upper section 18 of the reactor; however, BaSO ₄ particles which precipitate out are carried upwardly with the effluent. The upward liquid velocity in the upper section or second reaction zone 18, upstream of the take-off point of the line 26, is about 12m/h. Downstream of the take-off point of the line 26, the upward liquid velocity in the second reaction zone is about 5m/h. The BaSO ₄ particle containing effluent is withdrawn along the line 28 for further separation into BaSO ₄ sludge and effluent in the separator 30. The upward liquid velocity in the separator 30 is less than 3m/h. Clarification occurs in the upper section 18 where there is a reduction in upflow liquid velocity as compared to the lower section 14 because of the increase in the column diameter and the withdrawal of the recycle line 26.

Some (or even all) of the BaSO ₄ particles withdrawn as a sludge from the separator 30 along the line 32, can be recycled to the first reaction zone 14, e.g. along the line 24 (not shown).

To reduce the feed water calcium concentration to between 200mg Call and 400mg Call, a recycle water stream is withdrawn along the line 26 and reintroduced into the bottom of the reactor via the line 24. This results in reduction of the calcium carbonate precipitation potential in the lower section of the reactor and therefore the degree of supersaturation of the feed water so that crystal growth rather than nucleation of the calcium carbonate takes place, as hereinbefore described.

A CaCU3 sludge comprising CaCCb pellets, some residual water, and some barium carbonate and barium sulphate particles is withdrawn from the bottom of the reactor 12 along the line 36.

Typically, the CaCU3 sludge is withdrawn on a non-continuous or intermittent basis, while the withdrawal of effluent containing precipitated BaSO ₄ particles is effected continuously along the line 28. EXAMPLE 1 - Laboratory Scale

The method of the invention was assessed on laboratory scale as described in more detail hereunder.

Laboratory reactor design and operation

Column design

The laboratory reactor was manufactured from a 32mm clear PVC pipe (ID = 26.5mm). The pipe was made up of 1 m long sections which were coupled together using PVC couplers. This allowed the height of the reactor to be changed by inserting or removing pipe sections.

The feed water, barium carbonate slurry, carbon dioxide water and recycle streams were fed into the bottom of the reactor. Each stream was pumped using a Watson-Marlow (trade mark) peristaltic pump capable of pumping between 1.3 and 160£/h as controlled by changing the pump head tubes and the pump head speed using variable speed drives.

Stones, bolts and nuts were placed in the bottom of the reactor to mix the different streams and to disperse the reaction medium comprising the various streams across the reactor cross-sectional area. Approximately 200mm of the reactor height was filled with this large material, which thus dispersed the liquid streams as they entered the reactor. The remainder of the reactor was filled with calcium carbonate pellets sized between 0.4mm and 0.5mm. At the top of the reactor, the recycle water stream was drawn off; after a further short section of pipe above the recycle stream drawn off, a clarifier was fitted. This clarifier was initially made from a 320mm plastic funnel capable of retaining all of the particulate matter produced in the column, but later 50mm and 63mm piping was used, thereby providing a small increase in the column diameter and hence a reduction in the upflow liquid velocity. This is discussed further below.

Carbon dioxide water dosing

Carbon dioxide water used to neutralise the pH in the column was prepared by bubbling carbon dioxide gas through an upright 20mm PVC pipe. A container at the top of the pipe was kept full of water to maintain a constant contact time between the gas and water. The water was pumped from the pipe below the carbon dioxide dosing point. The gas flow rate, pipe length and liquid flow rate were changed so that the carbon dioxide acid load could be satisfied for the column. However, this system was not capable of dosing the small quantities of carbon dioxide required by the column, and consistently lead to a drop in the operating pH of the column to below pH 8. To reduce the carbon dioxide dosing rate, the gas was then bubbled through the barium carbonate feed slurry. The rate at which this gas was bubbled was varied to maintain a constant pH in the system. However, the lag time for this control system was extremely long with the pH only stabilising after several hours. Feed water (at p H 11)

A pilot plant ("A") is situated at Harmony (or Rand Uranium) No 8 Shaft in Tweelopies Road, Randfontein on the West Rand in South Africa. At this site, acid mine drainage is collected from several sources and is treated with lime for neutralization. AMD from the BRI dam where AMD ferrous concentrations are in excess of 1 ,200mg Fell was initially used, but supply problems were experienced. A constant supply of water from the No 8 Shaft was thereafter arranged, and this water was continuously supplied to the pilot plant site. This water was treated in the ferrite pilot plant to remove iron and manganese, and the pH was then raised to 11 using lime to remove soluble magnesium. The sulphate concentration in the treated water was close to the saturated gypsum concentration and varied between 1.4g and 2g SO ₄ /L

Barium carbonate feed

Three sources of barium carbonate were used. The first source was barium carbonate imported from China (Chinese barium carbonate, hereinafter referred to as CBC). The second source was prepared on site (South Africa barium carbonate, hereinafter referred to as SABC) by acidifying the CBC using hydrochloric acid, allowing the residual solids material to settle, and then adding sodium carbonate (also available on site) to precipitate the barium as barium carbonate. The final source was prepared in the laboratory using barium chloride and sodium carbonate, but only a small amount of this material was produced. The barium carbonate slurry water contained sodium chloride, which was not desirable for the study, and therefore several washing steps were performed until the conductivity of the water was similar to the conductivity of the water used to wash the slurry.

Recycle stream

For calcium carbonate pellet formation, it is necessary to reduce the calcium carbonate precipitation potential by reducing the concentration of the reactants at the bottom of the column. To achieve this, a recycle stream was introduced so that the calcium concentration at the bottom of the column could be reduced to below 350mg CaIL This was shown to reduce the calcium carbonate nucleation as measured by a reduction in the total calcium concentration present in the recycle stream. Analytical procedures

Sulphate

Aqueous sulphate was measured on the feed water (feed sulphate) and the effluent water. No total sulphate measurements were performed. The samples were filtered through a 0.45μm membrane filter paper to remove all solid material. Each sample was conditioned using a glycerine conditioning agent, barium chloride was added as a powder, and after mixing and standing, the turbidity was measured using a turbidity meter. This turbidity reading was compared with a calibration curve to determine the sulphate concentration.

Aqueous carbonate species

The total aqueous carbonate species were measured by titrating the samples with a standard hydrochloric acid solution. Effluent samples were collected and filtered through a 0.45μm membrane filter paper to remove all particulate material. For carbon dioxide water, several drops (exact number determined regularly) of concentrated sodium hydroxide were added to a volumetric flask. The carbon dioxide water was collected from the bottom of the absorption column directly into the hydroxide solution. Sufficient hydroxide was added so that the final pH of the samples was greater than pH 9.

The samples were titrated to a pH of around 3.7, and thereafter five further acid additions were made to a final pH of 2.7. The volume of acid used for this titration and the initial sample pH were then used to calculate the total alkalinity of the sample. From this calculation and the initial pH, the carbonate and hydroxide species were calculated, and therefore the total carbonate species concentration was calculated (as mg CaCCb/!!)-

Total carbonate species were not measured as these would have included the barium carbonate and calcium carbonate species. Metal analyses

The calcium, barium and magnesium concentrations were determined using an atomic absorption (AA) spectrophotometer. The preparation of the various samples is discussed in more detail below.

Barium

The total barium concentration was measured in the feed barium carbonate slurry, the effluent and the recycle. This barium was in the form of barium carbonate and dissolved barium. The barium carbonate was readily soluble in hydrochloric acid and therefore only acid addition, filtration and dilution was required in preparation for the AA analysis.

Aqueous barium was measured in the effluent from the column only. The sample was filtered and acidified.

Barium sulphate was considered insoluble at all times, and after acid addition, the samples retained a milky colour. This was removed through filtration. Because of the insolubility of the barium sulphate, it was not possible to calculate barium or sulphate mass balances for the system.

Calcium

The total and dissolved calcium were measured in the feed water since this water contained some solids carry over. The effluent water was analysed for total and soluble calcium. The total calcium would be in the form of calcium carbonate, while the dissolved calcium would be associated with either sulphate or in equilibrium with carbonate. The total calcium was measured in the recycle stream to determine the accumulation of calcium carbonate in this stream and therefore the system. The dissolved samples were prepared by filtration and acidification, while the total samples were first acidified and then filtered.

The total mass of calcium in the system could not be determined since it precipitates on the sand particles and other surfaces, and therefore a mass balance calculation could not be performed on the system. Magnesium

Total and dissolved magnesium was measured on the feed water and effluent streams, while total magnesium was measured on the recycle stream. Total magnesium samples were prepared by dissolving the magnesium hydroxide in acid before filtering. Soluble magnesium samples were prepared by filtering the sample before adding acid.

Results from laboratory-scale reactor

Recycle and clarifier investigation

The reactor was first operated as a flow through system where diluted feed water (at pH 1 1 ) was blended with barium carbonate slurry containing carbon dioxide. In this system, the aim was to determine the amount of sulphate removal in a single pass system. Stoichiometric quantities of sulphate and barium were added to the system, and operating parameters such as the barium source, operating pH and reactor height were changed to determine the effects.

The results showed that the Chinese barium carbonate was equally reactive in this system compared with the two other produced slurries. The results also showed that a lower pH (below 8) resulted in greater sulphate removal than higher pH, and that a change in reactor height had little effect on the sulphate removal efficiency.

However, in all of these tests, the maximum sulphate removal was only 700mg SO ₄ /I, with a barium sulphate conversion of only 35%.

The reactor was redesigned following a batch experiment which showed that if the contents were recycled for a further 15 minutes, complete sulphate removal could be achieved. The system was thus changed to include a clarifier at the top of the reactor constructed from a 320mm plastic funnel, and a recycle stream was included drawing the recycle water from the reactor below the clarifier (no solids settling or thickening). Operating the system with these additions showed that the CBC was closer to 70% reactive, while the SABC was 100% reactive, and the effluent sulphate concentration was controlled to less than 20mg SO4/L

The operating pH had little effect on the reactivity of the SABC and low effluent sulphate concentrations were maintained with an operating pH from 8.7 to 10.8. Also, the reactor height had little effect on the barium carbonate reactivity.

However, all of these parameters had an effect on the calcium carbonate precipitation, and the reactor had been designed with this precipitation in mind. At lower operating pH (below 9.2), the effluent contained higher concentrations of total carbonate species and soluble calcium, indicating that the low pH reduced the driving force for calcium carbonate precipitation. This presents the lower limit for the operating pH. The upper limit is determined by the final acceptable pH of the water rather than the kinetics of the reactor, and therefore a pH of between 9.2 and 9.6 was selected since this was slightly higher than the minimum for the reactor, but still acceptable for the effluent water. When operating the reactor at this pH, the effluent contained total carbonate species (C _T ) concentrations of 50mg CaCCb/!!, and when considering the load of carbonate to the system from both barium carbonate and carbon dioxide, the carbonate removal efficiency was high. This occurred when the reactor height was greater than 5.6m.

The final development saw a change in the diameter and therefore upflow velocity in the clarifier at the top of the column. When using the 320mm funnel, the effluent was almost solids free, indicating that even the barium sulphate was accumulating in the system. This led to the liquid in the system taking on the appearance of an emulsion because of the high fine solids load. The column diameter was 32mm (26.5mm ID), with an upflow velocity of 25m/h. In order to reduce the total solids accumulation being experienced in the column, the 320mm conical clarifier was replaced with a 63mm diameter pipe, so that the upflow velocity was reduced from 28m/h to 6.5m/h. Downstream of the recycle stream take off point, the upflow velocity was reduced further to 4m/h. This system allowed for good conversion of the barium carbonate while reducing the solids load in the system. Following this successful change, the clarifier diameter was reduced further to a 50mm pipe. In this pipe, the upflow velocity was reduced from 28m/h in the column to 10.1 m/h in the clarifier section to 6.3m/h above the recycle off take point. In this system, the effluent sulphate concentration remained similar to the 63mm clarifier experiment, indicating that there was no detrimental effect to the column by reducing the final clarifier diameter to 50mm.

This concluded the laboratory-scale development of the column, and the results obtained were used to design and operate the pilot-scale reactor described in Example 2. EXAMPLE 2 - First Confidential Pilot Scale Testing

The method according to the invention was then assessed on pilot scale as described in more detail hereunder. The same analytical procedures as described hereinbefore with reference to Example 1 , were used in Example 2. Pilot-scale development

Column reactor design

The main restriction governing the design of the pilot plant was the availability of feed barium carbonate. For this reactor or column, SABC was used, and several batches of CBC were processed to produce sufficient SABC for one week of operation. Therefore, the column was designed to treat 80t/h of feed water at pH 1 1.

The column was constructed from a 1 10mm clear PVC pipe. The length of pipe between the feed points and the recycle off-take point was 5.33m. The distance between the recycle off-take point and the effluent point was 0.43m. The 110mm clear PVC pipe made up 4.9m of the total height, ie constituted the lower section of the reactor. At this point, the column diameter was increased from 110mm to 160mm. This reduced the upflow velocity by more than 50% from 28.6m/h in the lower section of the column to 12.6m/h in the lower clarifier portion or intermediate reactor section. After the recycle off take point, the upflow velocity was reduced further to 5m/h by reducing the total flow rate through this section of the column. This allowed for settling of barium carbonate and other larger particulate material while not allowing for the accumulation of barium sulphate precipitate.

At the bottom of the column, the four feed streams (feed water, CO ₂ water, BaCU ₃ slurry, and recycle water) entered the column at slightly different points to reduce the formation of calcium carbonate scale. This will not be applicable to a full-scale process since the feed and recycle streams will be blended before entering the column and the configuration will not resemble the pilot column. The bottom of the column was filled with glass marbles to act as a distribution system for the liquid streams. The two main flows were the feed water and the recycle stream, and these two entry points were situated directly opposite each other on the column wall, so that the two streams collided with each other upon entering the column.

Screened sand with a particle size smaller than 0.75mm was initially introduced into the column, and was fluidized under the influence of upflow liquid velocity in the column. The minimum particle size was not known, as the sand was the waste material from the production of filter sand, where only larger particle sizes are required. No further particle size analysis was performed on the sand, but sampling can be carried out if necessary. The stable fluidised bed height made up 3.7m of the clear pipe length, corresponding to 76% of the column height with a contact time of 7 minutes 47 seconds.

Feed stream preparation

Feed water (at pH 11)

The ferrite pilot plant operated continuously for the week that the pilot column reactor was operated. In this system, the metals were removed by adding ferrite seed to the mine water at a pH of 9. This water entered a clarifier where all of the solids were removed. The clear water flowed into a mixed tank where lime slurry was dosed to maintain a pH of 11. This was the selected pH to remove magnesium from the mine water, and the magnesium concentration fed to the column, leaving the column and present in the recycle stream of the column was monitored as part of the plant operation. This is discussed further below. The feed water was fed into the reactor at 80UU.

Barium carbonate feed

SABC was produced in batches to satisfy the requirements for the pilot plant study. A 21Ot drum was filled with the correct blend of barium carbonate and tap water to allow the peristaltic feed pump to operate at a reasonable rate and to satisfy the barium load required for the process. A new batch of barium carbonate feed was made once per day from the concentration barium carbonate slurry stocks. Several samples of the feed barium carbonate were collected during the day but only analysed the following day, so that the actual load to the column was not known at any point. The barium carbonate slurry was fed at 5t/h.

The barium carbonate feed line had a tendency to block at the suction pipe when entering the peristaltic pump. This caused the effluent sulphate concentration to increase dramatically, but once the blockage was cleared, the system stabilised quickly. Again, this should not be a problem at full- scale, since the barium carbonate will not necessarily be fed to the system in this manner.

Carbon dioxide water dosing

The flow rate of carbon dioxide water was the smallest flow of any of the four streams being fed to the column. The carbon dioxide flow was used to control the pH of the column, but this flow needs to be extremely flexible in order to match the acid demand from the feed water. At a flow rate of 4.6<!/h, this line also tended to block with sand from inside the column, and when the effluent pH started to increase, this was typically the cause. Again, for a full- scale plant, this should not be a problem, since the carbon dioxide will not be dosed directly into the column, and therefore these blockages will be prevented. Also, by regulating the flow of carbon dioxide gas rather than saturated carbon dioxide water, the pH of the system could be controlled more accurately.

However, for this system, a 51 plastic bottle was used as a reservoir for storing the water. A toilet float valve was used to maintain the water level in this bottle. The carbon dioxide was dosed into a 20mm PVC pipe and allowed to dissolve into the water as the bubbles passed through the 1 m length of pipe.

The water was then dosed from the bottom of the vertical pipe from below the carbon dioxide dosing point. This allowed for a fairly consistent load of carbon dioxide to be dosed to the system with time, as indicated by the fairly constant effluent pH as shown below.

Recycle stream

The recycle stream was drawn off from the clarifier section at the top of the column. The recycle flow rate was maintained at 130t/h, resulting in a total feed dilution of (220/80) = 2.75 times. The recycle stream was analysed regularly for total calcium, barium and magnesium, since these three precipitates could accumulate in the system, and any accumulation would be evident in the recycle stream.

Column operation and performance

The column was operated for three consecutive days. The plan was to operate for four days, but several problems were experienced on the first night, which meant the system needed to be restarted the following day, and one day of operation was lost. However, three separate sets of operating conditions were tested with some useful results. Unfortunately, mass balances could not be calculated for the system, since it is not possible to measure the barium sulphate concentration in the effluent, and it is also not possible to measure a change in the mass of calcium carbonate in the column over a short period of operation.

Feed water quality

Figure 2 shows the pH and sulphate concentration of the pH 11 water as fed to the column. From Figure 2, the feed sulphate concentration was generally between 1 ,400 and 1 ,800mgSO ₄ /<!. The major variations to these average concentrations are probably a result of contamination of the sample or analytical error, since the sample was collected from the overflow of the clarifier with a retention time of several hours, and it would not be possible to change the sulphate concentration in the clarifier significantly, especially to reduce the concentration to below the saturation concentration of calcium sulphate.

Maintaining a constant pH of the feed water was a difficult exercise. Small increases in pH at a pH of 11 have a massive influence on the acidity demand and therefore the carbon dioxide dosing rate. The influence of the feed pH on the effluent pH is shown in Figure 5 below. The main problem is that the pH probe in the feed water tank was prone to scaling, and therefore the set point of the controller drifted in time. For a full-scale plant, the maintenance of the pH probes for this system will be a critical operating practice.

Carbon dioxide dosing

The carbon dioxide water quality is shown in Figure 3. From Figure 3, the pH of the carbon dioxide dosing water was fairly constant over the three days that the system was analysed. The total carbonate concentration was however highly variable. This is probably because of poor sample preparation, since the carbon dioxide water is super-saturated with carbon dioxide at atmospheric concentrations, and losses of carbon dioxide will occur immediately once the sample has been collected.

Effluent sulphate concentration and pH

The effluent sulphate concentration and pH was measured hourly for the duration of the pilot plant operation. The results of this measurement are shown in Figure 4. From Figure 4, three different operating periods can be defined. For the first period, the effluent sulphate concentration was constant at around 500mgSO ₄ /t. For the second period, the concentration dropped to around 200mgSO ₄ /t, while for the final period, this dropped further to less than 100mgSO ₄ /t. For each of these periods, the feed streams, effluent stream and recycle stream were analysed in detail, and the results from the three periods are discussed in more detail below. Figure 5 shows the influence of the actual pH of the feed water on the effluent pH. During these periods of variable feed water pH, the carbon dioxide dosing was not changed, and therefore the same acid load was being applied to the system. Clearly, for a full-scale operation, the range of carbon dioxide dosing will need to be significant to maintain a stable operating pH in the column.

From Figure 5, when the feed water was maintained within 0.3 pH points from pH 11 , the effluent pH was within the expected range of 9.2 to 9.8. However, when the feed water pH increased to more than pH 1 1.5, the effluent pH increased to above 10.5, and with a further increase in the feed water pH to close to pH 12, the effluent pH increased to more than pH 11. The logarithmic pH scale requires that the carbon dioxide dosing rate be capable of dosing accurate quantities of carbon dioxide over a range of several orders of magnitude, and the whole system relies on the accuracy of the pH measurement.

Table 1 : Summary of the pilot plant operating data

Three periods of constant operation

Table 1 shows the average data from the three periods of steady but different operation as described above. The three periods were defined by the different effluent sulphate concentrations. The data presented in Table 1 shows that the different effluent sulphate concentrations were a direct consequence of the different barium loads to the column. Table 1 shows that the fraction of sulphate removed for each period (68.25%, 84.90% and 96.39% for periods 1 , 2 and 3 respectively) corresponds to the barium to sulphate loading ratios of 61.5%, 78.3% and 102.8% respectively. Therefore, by adjusting the barium carbonate load to the system, it is possible to manipulate the effluent sulphate concentration. From the barium loading data, the barium concentration is measured by dissolving the barium carbonate in acid and measuring the particulate and dissolved barium concentrations using an atomic absorption spectrophotometer. Therefore, the total barium concentration contains both soluble barium and unreacted barium carbonate. It is assumed that barium sulphate will not dissolve in this acidic solution. The results shown in Table 1 indicate that the barium carbonate is readily consumed in the system, and when it is not dosed in excess as in Period 3, less than 2% of the barium remains unreacted. The effluent soluble calcium concentration is directly linked to the effluent sulphate concentration, and the results in Table 1 show a fairly strong correlation to this. The problem is that there is a poor correlation with the total and soluble effluent calcium concentrations. The aim of the system is to precipitate calcium as calcium carbonate on the existing calcium carbonate pellets. It is not possible to determine the effluent particulate calcium carbonate concentration from the data present, and therefore the efficiency of the system in removing calcium carbonate as pellets. However, if calcium carbonate was not precipitating as pellets, the effluent total calcium concentration would be significantly higher if the calcium carbonate precipitate was not accumulating in the system. However, based on the total calcium concentration present in the recycle, it is unlikely that this is the case. Therefore, it seems that calcium carbonate is indeed precipitating on the pellets in the column.

The results of the magnesium analyses show that the secondary lime treatment stage is not capable of removing all of the soluble magnesium from the mine water, and it may be necessary to operate the pH 11 system at a higher pH in order to achieve complete removal.

EXAMPLE 3 - Second Confidential Pilot Plant Testing

Introduction

At the beginning of 2009, a confidential pilot plant in accordance with the invention ('ABC Process') consisted of the unit operations as indicated in Figure 6.

In Figure 6, reference numeral 50 generally indicates a pilot plant installation for carrying out, on pilot plant scale, the method according to the invention for treating calcium, sulphate and hydroxide ion containing water.

The installation 50 includes a first treatment stage 52 with an acid mine drainage or impure water line 54 leading into the stage 52, as does a lime addition line 56. A metal sludge and gypsum withdrawal line 58 leads from the stage 52, as does a water transfer line 60. The water pH is raised to about 9 in the stage 52, to precipitate metals and sypsum.

The line 60 leads into a second treatment stage 62, as does a lime addition line 64. A magnesium hydroxide and gypsum withdrawal line 66 leads from the stage 62, as does a water transfer line 66. The water pH is raised to about 11 in the stage 62, to precipitate magnesium hydroxide and gypsum.

The line 66 leads into a third treatment stage or reactor 70, as does a carbon dioxide addition line 72. A calcium carbonate withdrawal line 74 leads from the stage 70, as does a water transfer line 76. The water pH, in the stage 70, decreases to about 9.

The line 76 leads into a fourth treatment stage or reactor 80, as does a barium carbonate addition line 82. A calcium carbonate and barium sulphate withdrawal line 84 leads from the stage 80 as does a water withdrawal line 86.

The line 84 leads into a fifth treatment stage 90 which constitutes a thermal reduction and processing stage. A barium carbonate withdrawal line 92 leads from the stage 90, for recycling barium carbonate to the stage 80 (line 82). An ash withdrawal line 94 leads from the stage 90, as does an elemental surphur withdrawal line 96.

The operation of the pilot plant installation 50 is similar to that of the installation 10, and also emerges from what is set out hereunder.

Since the beginning of 2009, the carbon dioxide addition (stage 70) and barium carbonate addition (stage 80) were consolidated in one, i.e. the same, reactor (not shown). A current modification (not shown) is the combination of a pH correction reactor (stage 70), where carbon dioxide is dosed to produce calcium carbonate, and a sulphate removal stage (stage 80), where barium carbonate is dosed to precipitate calcium carbonate and barium sulphate. Three different reactor configurations were evaluated for this combination of processes in a single reactor.

The first reactor configuration was a mixed tank with sludge recycle in the classic contact reactor configuration, and was labelled the "A" pilot plant. The second configuration recycled the sludge internally to improve solids contact and therefore crystal growth, and was labelled the "B" pilot plant. The third pilot plant aimed to separate the barium sulphate sludge from the calcium carbonate sludge by producing calcium carbonate pellets, and this was labelled the pellet reactor, and was in accordance with the invention. The pellet reactor was operated for a short period in September 2009, while all three of the different reactor configurations were operated between March and May 2010 at a confidential pilot plant facility. Mine water was used as feed water, and all three systems were operated using pretreated mine water and barium carbonate produced on site. Therefore, the performance of each of the plants is comparable. The design and operating parameters of the three different pilot plants is discussed hereunder.

Reactor design and operation

(a) Barium sulphate and calcium carbonate reactor (combination of reactors 70 and 80 in Figure 6)

In this reactor, water containing super-saturated calcium sulphate together with high concentrations of dissolved lime is treated by adding carbon dioxide gas and barium carbonate. The saturated calcium sulphate reacts stoichiometrically with the barium carbonate to precipitate calcium carbonate and barium sulphate (Reaction (1 )). The high concentration of lime present due to the high pH is neutralised and precipitated by the addition of carbon dioxide to form calcium carbonate (Reaction (2)). (b) Configurations

As indicated hereinbefore, three reactor configurations were operated at pilot scale in Example 3, viz "A", "B" and the pellet reactor. The pellet reactor was developed for the separation of the calcium carbonate sludge from the barium sulphate sludge.

(c) Pellet reactor

A sketch of this reactor is shown in Figure 7 in which reference numeral 100 generally indicates the reactor. The reactor 100 consisted of a 400mm ID uPVC pipe 102 which was 5m in height. Below this pipe, another section 104 of 350mm ID uPVC pipe was fitted. All of the feed streams entering the reactor, entered the section 104. Thus, a 32mm PVC recycle pipe 106 entered the section 104 250mm from the bottom of the reactor; a 50mm PVC sludge drainage pipe 108 led from the section 104 near the bottom of the reactor; and a 25mm PVC water feed pipe 1 10 entered the section 104 400mm from the reactor bottom. The upflow velocity in this section 104 was higher than in the bulk 102 of the column, and this prevented smaller particles from settling in this zone. At the top of the reactor 100, a 537mm ID pipe 112 (length 0.9m) was fitted to act as a clarifier. The sludge and water were recycled from this clarifier back to the bottom of the column (pipe 106). Above the recycle off-take point, the low upflow velocity allowed for settling of the heavier or larger solid material, with only fine particles leaving with the over flow water (pipe 114). These solids were collected in a separate clarifier.

The upflow velocity at the bottom of the column was more than 50 m/h, while for the bulk of the column, it was 25 m/h. At the top of the column, in the clarifier section before the recycle off-take point, the upflow velocity was 12 m/h, and above the recycle off-take point, it was 5 m/h. The upflow velocity in a final clarifier (not shown in Figures 6 and 7) which stood next to the column was less than 0.5 m/h.

(d) Mine water preparation

For all of the pilot plant testing in Example 3, mine water was pretreated before being dosed to the sulphate removal process. The water was pumped into a 5,00Ot mixed tank, where hydrated lime was added to maintain a pH of 11.2. The retention time in this tank varied from 3 to 10 hours, depending on the pilot plants that were being operated. A 10% lime slurry was prepared in a 2,50Ot make-up tank using fresh water, and this was pumped periodically to a 50Ot dosing tank.

The water then flowed to a clarifier, from where the solids were pumped to sludge holding tanks. The clear water over flow from the clarifier flowed into a holding tank, from where it was pumped directly to the pilot plant reactors. An over flow point was installed into this tank so that excess water could flow to a drain. The sulphate, calcium, magnesium and iron concentrations and the pH and conductivity were measured regularly on samples collected from the holding tank. (e) Barium carbonate preparation

Barium carbonate was produced by dissolving barium sulphide in water, separating (settling) the solid material, and precipitating the dissolved barium sulfide as barium carbonate by adding sodium carbonate. The resultant solution contained barium carbonate precipitate and dissolved sodium sulphide. The slurry was washed with tap water in a counter-flow column to remove the sodium sulphide and to produce clean barium carbonate for dosing to the barium sulphate reactors.

The processes were designed to produce a barium concentration of 100gBa/t, although difficulty was experienced in achieving this.

For the first period of operation of the pellet reactor, CBC was used. This material was shown (see Example 1 ) to be less reactive than freshly produced barium carbonate. This was thus known, but since no other source of barium carbonate was available, the process was started on this material.

(f) Pellet reactor

The pH 11.2 water and recycle water were fed into the narrow section at the bottom of the column. The pH 11.2 water was fed at a rate of 1 ,OOOt/h, while the recycle flow rate was 1 ,600t/h. The barium carbonate and carbon dioxide were dosed into the recycle line. The pH was measured at the top of the clarifier section, and the carbon dioxide was controlled manually according to the measured pH. The system was operated from 9 until 30 September 2009, and again from 15 April until 25 May 2010. (g) Analytical procedures

Sulphate

This was done in the same manner as in Example 1 , with aqueous sulphate being measured on the pH 11.2 water (feed sulphate) and effluent water.

Aqueous carbonate species

This was also done as in Example 1. Metal analyses

These were done using the procedures described in Eample 1 ; in all cases the feed water was the pH 11.2 water.

Pilot plant results

(a) pH 1 1.2 water preparation

The aim of adding lime to a pH of 11.2 is to precipitate calcium sulphate (gypsum, CaSO ₄ .2H ₂ O), iron as ferrous hydroxide complexes, manganese and magnesium. The removal of metals from the mine water would be almost complete at this pH, while the precipitation of gypsum would tend toward saturation concentrations.

Since the systems were operated over two periods, the pH 11.2 water was produced over two periods, and these are shown in Figures 8 and 9 for the September 2009 and March to May 2010 operating periods respectively.

From Figures 8 and 9, the average sulphate concentrations for the two periods were 2394 and 2418mgSO ₄ /t, showing that by raising the pH to 11.2 using lime, the sulphate concentration can be controlled. This obviously depends on the operation of the pH 1 1.2 reactor, since the average sulphate concentrations are higher that the solubility concentrations for gypsum and the actual sulphate concentration in the effluent from the pH 1 1.2 reactor may vary according to the retention time, solids concentration and other ions present in the mine water. The pH of the pH 11.2 reactor was controlled by a pH controller connected to a dosing pump, and this allowed the pH to be tightly controlled. The actual pH was measured and recorded every hour, with an average pH of 1 1.21 ± 0.02 (1236 points). The conductivity of the pH 1 1.2 water was also measured every hour, with an average of 3.84 ± 0.32 (506 points). The significance of this will be discussed later when the conductivity of the effluent is shown.

During the second operating period, the iron and magnesium in the pH 1 1.2 water was measured regularly. The average aqueous iron concentration was 0.14 ± 0.74mgFe/<! (231 points), while the average magnesium concentration was 0.51 ± 0.69mgMg/t (214 points). This shows that the pH 1 1.2 reactor was efficient in removing these two metals from the mine water feed. The total iron and magnesium concentrations measured in the pH 1 1.2 water were 6.11 ± 31.76mgFe/e (229 points) and 2.09 ± 9.30mgMg/t (210 points), indicating that the clarifier was at times unable to cope with the solids load.

Figure 10 solution: An alternate method was finally used where barium sulphide was dissolved in water, the solid material was separated (settled), and the barium precipitated as barium carbonate by adding sodium carbonate. The resultant solution contained barium carbonate precipitate and dissolved sodium sulphide. The slurry was washed with tap water in a counter-flow column to remove the sodium sulphide and allow for clean barium carbonate to be dosed to the barium sulphate reactors. (b) Pellet reactor during the first operating period

During September 2009, the pellet reactor was operated. CBC was dosed into the system, since locally produced barium carbonate was not available.

Figure 10 shows the sulphate and calcium concentrations in the effluent from the system. From Figure 10, the sulphate and calcium concentrations varied considerably in the initial stages of the operation (first 15 days), but the sulphate and calcium removal was almost complete for the last 10 days of operation. This is due to the operation of the barium carbonate dosing pump. Initially, this pump was prone to blocking and the operators were not diligent in checking the dosing pump, but once this had been resolved, the performance of the plant improved considerably.

As mentioned hereinbefore, the CBC was less than 100% reactive, and therefore it was likely that barium carbonate was building up in the system. Once the barium dosing had been rectified, it was possible to achieve more than 99% sulphate removal, as was witnessed during the earlier development of this reactor. The calcium concentration dropped to less than 5mgCa/<! at the end of the operating period, which corresponded to the removal of the sulphate, indicating that with sufficient barium carbonate addition, both calcium carbonate and barium sulphate precipitation can be achieved. The average aqueous barium concentration for this operating period was 3.35 ± 1.61 mgBa/<! (21 points), indicating that some barium is present in the aqueous phase.

Figure 11 shows the relationship between the sulphate concentration, pH and conductivity of the effluent from the system. The pH and conductivity were not measured for the second portion of the operating period, but from the data collected, the conductivity follows the trend of the sulphate concentration, but is not influenced by the pH of the water. The pH was difficult to control using the manual system that was in place, since the operators found it difficult to make small adjustments to the carbon dioxide dosing rate, and would tend to switch the dosing either on and off depending on the measured pH, rather than making small adjustments to maintain the pH between two set points. This was rectified later in the project, but should not be a problem with a larger scale operation, where on-line pH control will take place. No analysis of the solids took place during this period, and the operation was stopped once the barium carbonate supply was exhausted. (c) Pellet reactor

The pellet reactor was operated for the second term from 15 April until 25 May 2010.

Figure 12 shows the effluent sulphate and calcium concentrations for the pellet reactor. As for the reactor clarifier, the calcium and sulphate concentrations follow similar trends, indicating that this was mostly influenced by the barium carbonate addition.

Figure 13 shows that the effluent sulphate concentration closely followed the inverse of the barium feed concentration, while Figure 14 shows that the calcium and C _T concentrations seem to be independent of each other. These trends are all similar to that experienced by the reactor/clarifier, and the likely reasons for these trends have been discussed above. The main aim of the pellet reactor was to separate the calcium carbonate produced from the barium sulphate precipitate, by forming calcium carbonate pellets and fine barium sulphate particles. This separation will be discussed in the following chapter. (d) Summary

The following can be summarised from the results obtained in the pilot plant studies of Example 3:

• The pellet reactor was operated successfully with a retention time of 38 minutes and a recycle to feed ratio of 1.625

• In all systems, the barium carbonate feed concentration was lower than expected, and the systems were unable to remove the sulphate consistently to the desired concentration of 200mgSO ₄ /<! because of the barium concentration. However, when barium concentrations of 100gBa/<! were dosed, the effluent sulphate concentration was less than 200mgSO ₄ /<! • Although it is not obvious from the data, and therefore not presented as such, it should be possible to control the pH and therefore calcium concentration by controlling the carbon dioxide dosing rate, and to control the sulphate by controlling the conductivity of the water, and on-line control of both of these parameters is practical

• All of the above results may need to be verified using the correct source of barium carbonate, since this could significantly influence the performance of the reactors

The following table (Table 2) gives an indication of the mass balances that could be calculated from the data. The data presented previously in this chapter shows the variations experiences in the daily performance of the systems, and should be a caution against the absolute values included in the summary. However, some important values are calculated which may assist in designing and operating a full-scale plant.

From Table 2, the sulphate removal loads are similar for all systems (70-80%), and this is due to the barium load to the system, which was between 17 and 25% less than required. Although the effluent contained some particulate and aqueous barium, this was a relatively small fraction of the total barium load (0.3-1.4%). All three systems showed similar particulate calcium removal efficiencies (79-85%), and this could have been improved by increasing the clarifier size for each process. In fact, for all systems, and increase in the clarifier size would be recommended, although for the pellet reactor, the clarification section above the fluidised bed section is adequate. The additional clarifier required to remove barium sulphate from solution was not investigated in this study. Barium sulphate clarification design equations need to be developed before a full-scale plant is designed. Table 2: Summary of operating data and reactor performance

Pilot plant "A" "B" Pellet

Aqueous sulphate load in (g/h) 1223 1220 1966

Aqueous sulphate load out (g/h) 256 278 549

Sulphate removal 966 941 1418

% Sulphate removal 79% 77% 72%

Required barium load (g/h) 1745 1741 2806

Total barium load in (g/h) 1449 1306 2329

Equivalent sulphate load (g/h) 1015 915 1632

% Barium shortfall 17% 25% 17%

Aqueous barium load out (g/h) 1.08 1.63 2.85

Particulate barium load out (g/h) 3.23 16.95 16.12

Total barium loss to effluent 0.3% 1.4% 0.8%

Total calcium load in (g/h) 737 556 881

Aqueous calcium load in (g/h) 441 478 764

Particulate calcium load in (g/h) 296 79 117

Aqueous calcium load out (g/h) 71 89 192

Particulate calcium load out (g/h) 142 63 106

Aqueous calcium removal (g/h) 369 388 572

Particulate calcium removal - - 85% efficiency

Sludge production and processing

(a) Particle size analysis

Sludge samples were collected from each of the three pilot plants for particle size analysis. Two samples were collected from the pellet reactor over flow, which is expected to be barium sulphate, and from the under flow, which is expected to be calcium carbonate pellets.

Some of particle size analysis data are shown in Table 3. From Table 3, all of the reactor configurations produced a sludge consisting of particles in the range of 3 to 26 μm, suggesting that this is the standard size of the mixture of barium sulphate and calcium carbonate as produced by the process. The only exception was the pellet reactor under flow, which produced particles of 431 μm diameter, which is the result of calcium carbonate precipitation on the sand particles.

Table 3: Particle size distribution results

Sample Peak size % Volume

20.46 96.2

"A" under flow

4.73 3.8

19.19 89.6

"A" under flow

4.65 10.4

26.22 47.5

"B" under flow 11.09 31.6

4.12 20.9

17.1 1 9.3

Pellet reactor over flow

3.06 90.7

23.43 47.7

11.29 12.0

Pellet reactor over flow

7.58 12.4

4.12 27.9

Pellet reactor under flow 431.3 100

(b) XRD Analysis

Sludge samples were collected from the reactor/clarifier and pellet reactors during May 2010 and submitted for XRD analysis. Two methods were used to determine the relevant compositions of the solids. The difference between the two methods results in a significant change in the calcium and silica concentrations, which obviously changes the calcium carbonate ratios in these systems, and was reported on as follows:

Due to phases with relatively large differences in mass absorption coefficients (i.e. Calcite and Barite), the quantification may not be as accurate as desired, all samples were however analysed using the same microabsorption parameters - as established in previous work (first table). The second table shows results where no microabsorption correction was applied and those represent the maximum Ca- and Si- containing phases quantities. The quantities of Ca- and Si- containing phases have to be confirmed by chemical analysis.

Table 4 shows two sets of results, referring to the first and second tables from the XRD analysis report. The upper set of data show that the calcium carbonate concentration in the over flow of the reactor/clarifier and pellet reactor systems was low (<6% of the total solids), with the remainder being barium sulphate. This result is misleading to some extent, since the over flow stream from the reactor/clarifier should not have contained any particulate material, while the over flow stream from the pellet reactor passed to a barium sulphate clarifier where this product was removed. From this result, and barium sulphate sludge of about 95% purity could be produced.

The underflow sludge from the reactor/clarifier contained less calcium carbonate than expected, but also contained unreacted barium carbonate. For the pellet reactor, no barium carbonate was measured, and the calcium carbonate composition was greater than for the reactor/clarifier.

Table 4: Results of XRD analysis

Reactor Sample Date CaCO ₃ BaSO ₄ BaCO ₃ Quartz

O/F 21 May 3.3% 96.7%

D

U/F 7 May 11.1% 67.3% 21.6%

O/F 14 May 4.3% 95.7%

Pellet O/F 21 May 5.8% 94.2%

U/F 7 May 24.3% 69.8% 5.9%

O/F 21 May 1 1.4% 88.6%

D

U/F 7 May 32.2% 52.2% 15.6%

O/F 14 May 14.3% 85.7%

Pellet O/F 21 May 18.3% 81.7%

U/F 7 May 48.3% 37.4% 14.3% Of significance is the large variation in the compositions depending on the microabsorption correction that was applied.

In accordance with the invention, feed water containing dissolved calcium sulphate is thus fed directly, with CO ₂ water, into a reactor so that a separate carbon dioxide dosing step and classification step are thus not necessary. Furthermore, the difference in settling velocities between the precipitated calcium carbonate, as pellets, and the precipitated barium sulphate, as a fine precipitate, allows for the separation of the two sludges, and the production of a purer barium sulphate sludge, as a direct consequence of the method. The avoidance of a separate carbon dioxide dosing step is also advantageous.

Furthermore, the higher concentration of calcium carbonate when in pellet form (as compared to known processes in which calcium carbonate merely co-precipitates with barium sulphate as hereinbefore described), enhances the rate of barium sulphate precipitation from a nominal retention time of more than 1 hour to less than 10 minutes.