CA1340279C

CA1340279C - Sintered high titanium agllomerates

Info

Publication number: CA1340279C
Application number: CA000606689A
Authority: CA
Inventors: John Sydney Hall; Ken George Carey; Michael John Hollitt
Original assignee: Commonwealth Scientific and Industrial Research Organization CSIRO
Current assignee: Commonwealth Scientific and Industrial Research Organization CSIRO
Priority date: 1988-07-26
Filing date: 1989-07-26
Publication date: 1998-12-22
Anticipated expiration: 2015-12-22
Also published as: EP0426731A1; ZA895676B; ATE105873T1; WO1990001073A1; KR0148343B1; KR900702058A; DE68915446T2; AU3989889A; WO1990001072A1; AU3989789A; AU626191B2; JP2779028B2; EP0426731B1; AU626155B2; EP0426731A4; OA09635A; RU2080396C1; KR900702059A; JPH04500984A; ZA895675B

Abstract

A process for increasing the particle size of fines of a titaniferous mineral containing more than 45% by weight of titanium. The process comprises mixing the fines with a binding agent and water to produce an agglomerate. The agglomerate is then dried and sintered.

Description

1~ 10279 The present invention relates to agglom-rates of titanium-bearing material suitable for producing TiC14.
In prior art processes, materials of high titanium dioxide content (about 85% by weight TiO2) are the preferred raw materials for TiC14 manu-facture, subject to specifications on the particle size of the materials and on the content of some impurity elements.
TiC14 is a low boiling liquid which may be purified by distillation and chemical methods, following which it may be burned in oxygen to generate TiO2 pigment and chlorine gas, or reacted with magnesium or electrolysed to produce titanium metal.
The raw material, a titanium-bearing mineral sized within the range 100 - 300 microns (um), is fed to a fluidised bed reactor where it undergoes reductive chlorination at temperatures in the range 900~ - 1000~C. Petroleum coke or a similar high fixed carbon material is added to the bed as both fuel and reducing agent. Oxygen may be added to the inlet stream to maintain reaction temperatures.
The product TiC14 passes from the reactor in a gaseous form together with the gaseous chlorides of impurity elements and entrained fine solid particles from the fluid bed. The gases are cleaned of solids and condensed. The product TiC14 is purified by distillation and chemical methods.
In the chlorination stage, most metallic impurities form volatile chlorides, which leave the reactor in the TiC14 gas stream. However, the alkali and alkaline earth metals form relatively non-B

volatile chlorides which are liquid at reaction temperatures and hence tend to form agglomerated masses in the bed to the point of potential shut down. Accordingly, operators of the process usually specify stringent limitations on the contents of these elements in raw materials.
Impurities such as iron represent an economic penalty to the process in that they consume coke for their reduction and, more importantly, expensive reagent chlorine which is lost in waste iron chlorides. Silicon and aluminum are also partly chlorinated in the process, causing excess chlorine consumption. Aluminum chlorides are also the source of corrosion problems in process equipment.
As a mineral particle is progressively chlorinated, it reduces in size until it reaches a point at which it is entrained in the gas stream and leaves the reactor as an unavoidable and irre-coverable loss. Conventially, entrainment losses may amount to 5 - 10% by weight, of the input materials.
As the feed size is reduced below 150 um in .. .. _, . _ ~ 4 ~ 13 10 273 diameter, entrainment losses become relatively much higher than for materials of the conventional size. Such losses are both economically and operationally acceptable.
In an attempt to overcome these difficulties, in one 5 process known in the prior art, fine-grained TiO2-bearing material for fluidised bed chlorination is prepared by coking into composite agglomerated particles a mixture of TiO2-bearing material, bituminous coking coal and a water soluble binder. This prior-art process, however, has not been 10 accepted by the industry. One reason is that the chlorination process is reductive chlorination and so the carbon in the feed material must be present in a specific proportion to the TiO2-bearing material which may not be suited to composite strength development. Further, the agglomerate, because the 15 carbon is attacked, breaks down before complete chlorination occurs and so fine particle size material is lost to the process through entrainment in the gas stream.
In another process described in the prior art, a water emulsion of asphalt is used as a binder in the formation 20 by extrusion of pellets of fine-grained titanium-bearing material. By a process of slow curing at 1000~C, water is removed from the pellets and the organic material converted to carbon. The curing results in the caking of the binder in the pores and around the grains, forming a good bond. There is no 25 chemical bond between the binder and the titanium-bearing material. The extruded material must be broken before curing into a size range close to the required product size. This removes the need for the circulation of cured fines which would otherwise reduce the strength of the product pellets.
30 During chlorination of the pellets, the carbon takes part in the reductive chlorination process. This product therefore suffers from the same disadvantages as those described in the previous example of the prior art.
It is an object of the present invention to overcome 35 one or more of the perceived difficulties in the fluid bed chlorination of fine grained titanium-bearing materials.

-- 5 - 13;10~73 Accordingly, the present invention provides a process for increasing the particle size of fines of a titaniferous mineral containing more than 45% by weight titanium which process comprises:
mixing the fines with a binding agent and water to produce an agglomerate, drying the agglomerate and sintering it.
The agglomerated particles so formed are resistant to degradation forces associated with transport and handling.
10 The agglomerated particles are also resistant to the physical and chemical degradation forces and temperatures associated with chlorination processing including fluidised bed reductive chlorination processing.
The agglomerated particles, may be manufactured to 15 fall within a preferred size range to suit the dynamic requirements of fluidised bed reductive chlorination processing for example between 100 - 500 ~m, more preferably from approximately 150 - 250 ym. If particles fall below this range they may be entrained in the gas stream and therefore 20 lost to the reaction. If particles fall above this range they may cease to be buoyant within the fluidised bed and form an inactive layer at the bottom of the reactor.
The titanium-containing particles may be of any suitable titanium-containing mineral or minerals. The 25 titanium-containing minerals may be natural or synthetic in origin. The titanium-containing mineral may be a detrital mineral. The titanium may be present in the titanium-containing minerals in the form of titanium dioxide.
The titanium dioxide content of the titanium-containing 30 minerals may be approximately 85% by weight or greater. A
preferred titanium dioxide containing source is a deposit which includes any of the minerals rutile, anatase and leucoxene.

- 6 - t~40279 The titanium-containing minerals may be subjected to initial concentration processing after extraction. Initial concentration processing may increase the average titanium dioxide content for example to approximately 90% by weight or 5 above.
One titanium-containing mineral deposit at Horsham, Victoria, Australia of this type is further characterized by usually fine sizing. The unusually fine sizing suggests that major entrainment losses may ensue from later treatment by 10 reductive chlorination in a fluid bed.
The titanium-containing mineral may be present in any suitable amount in the agglomerated particles. The titanium-containing minerals may be present in amounts of approximately 95 - 99.5% by weight based on the total weight 15 of the sintered agglomerate.
The amount of water added may vary depending upon the size distribution of the original titanium-containing particles and the required size of the agglomerates. The amount of water may vary from approximately 5 to 15% by 20 weight, preferably approximately 8% by weight, based on the total weight of titanium-containing particles, binder and water.
The binder or binders for the titanium-containing particles may be of any suitable type. The binder for the 25 titanium-containing particles should be such as to form agglomerates capable of withstanding the physical, chemical and thermal degradation forces in the drying and firing stages of the process. The binder may be an organic or inorganic binder. The binder may be a ceramic or glass-forming binder.
30 The binder may be a carbon-free binder. A single binder may be used. A combination of two or more binders may be used to provide strength under the different operating environments of the drying and firing stages.
Calcium- or sodium-containing binders are not 35 preferred. This is so since the calcium or sodium contents of the binder may react in the reductive chlorination process to _ ._ . ... ~ .. .. . .. . _ . . . . .. ... , _.__ ~ 7 ~ 1~0~79 form deleterious liquid residues. Binders may contain calcium or sodium but should not result in the addition of these elements to cause problems in chlorination.
The binder for the titanium-containing minerals may 5 be such that it does not seriously contaminate the bound titanium-bearing particles for subsequent processing, for example in reductive chlorination processing.
The binder for the titanium-containing particles may include:
1) Colloidal silica 2) Silica, water soluble silicates or silica/fluorite mixtures 3) Clay minerals (including bentonite, kaolinite and montmorillonite) 4) Boehmite 5) Boehmite/silica mixture 6) Geothite 7) Lignosulphonate 8) Sodium carbonate (saturated water solution) 9) Sodium silicate 10) Group II element carbonate/clay mineral mixture 11) Sugars e.g. molasses 12) Aluminium salt/organic amide mixtures 13) Titanium bearing organic and inorganic solutions 14) Polyvinyl acetate 15) Water emulsified organic binders The amount of binder for titanium-containing 30 particles should be sufficient to produce a competent agglomerate. The amount of binder should preferably not be sufficient to encapsulate the titanium-containing particles.
A relatively low percentage of binder is preferred.
Percentages in the range of approximately 0.5 - 5% by weight 35 are preferred.

_. .

The mixing step in the process according to the present invention may be conducted in any suitable manner.
Agglomeration may be conducted in devices incorporating a rolling/tumbling action such as rotating disk or drum - 5 pelletisers or V-blenders, or in devices incorporating an impacting/shearing action such as high intensity micro-agglomerators or mixers, or in devices incorporating both actions. Agglomeration may be conducted in stages or in closed circuit with product sizing screens.
The drying step may be conducted at elevated temperatures e.g. 75 to 150~C. The drying step is preferably carried out in such a manner as to limit the residence time of the agglomerates in this part of the process to less than 30 minutes. The drying step may be conducted in any suitable 15 drying apparatus. A fluidised bed dryer or rotary dryer may be used.
In the firing step, the temperature and residence time should be sufficient to produce homogeneous or heterogenous phase bonding between the particles within the 20 agglomerates. The agglomerates may be heated to a temperature of approximately 1000~C to 1500~C preferably 1200~C to 1400~C.
The residence time of the agglomerates within the above temperature range may be for a period of approximately 5 minutes to approximately 6 hours.
The firing step may be carried out in any of a number of suitable means, including fluidised bed, oven or kiln firing.
In a preferred form of the present invention the process may include the preliminary step of grinding at least 30 a portion of the titanium-containing particle source.
The preliminary grinding step may be utilised to improve the size control in the preparation of the agglomerates and thus provide a greater strength and density to the fired product. The titanium particles may be 35 introduced into any suitable grinder. A ball mill or rod or intensive milling device may be used.

10~79 The amount of titanium-containing feed to be ground may vary from 0 to approximately 100% by weight depending on the source and type of titanium-containing material.
The grinding step may provide particles having an average size from approximately 1 ~um to approximately 50 um.
The sintered agglomerate may include a plurality of sintered agglomerated particles. The bond formed between the titanium-containing particles may include particle boundary recrystallization, that is, the boundaries of the titanium-containing particles may be physically merged. The bond formed between the titanium-containing particles may in addition include a bridging with a secondary phase formed by the binder. The sintering step may tend to reduce or eliminate the binder from the agglomerated particles. The initial binder may be burned off in whole or in part. The initial binder may be present and/or may be incorporated in whole or in part in the crystal lattice of the particles.
The present invention will now be more fully described with reference to the following non-limiting examples and the accompanying drawings, in which:
Fig. 1 is a diagram showing the fluidisation behaviour of WIM agglomerates; and Fig. 2 is a diagram showing the size distribution of products of fluid bed firing of titania agglomerates (25% fines).
B

- 9a -.
13 10~73 A laboratory scale bath Patterson-Kelley V-blender was used initially to blend a mixture of 9.2 kg of dry leucoxene with 1% by weight, of dry bentenite powder for 1 to 2 minutes. The leucoxene consisted of 75% in the size range 50 ,um - 10 ,um and 25% in the size range -50 ,um. The size distributions of the two fractions are recorded in Tables 1 and 2.

~ ~0~73 Size Distribution of Ground and Sized Leucoxene (-100 + 50~um) SIZE (ym) CUMULATIVE % PASSING
106 95.0 36.4 53 8.2 38 0.9 33 0.6 24 0.1 17 0.1 8 0.0 0.0 Size Distribution of -50ym Fraction of Ground Leucoxene SIZE (~um)CUMULATIVE % PASSING
106 100.0 99.6 53 99.1 38 90.4 33 88.6 24 59.6 17 33.2 8 10.1 0.0 The V-blender rotated at a speed of 40 rpm. Water was then introduced into the mixture through an intensifier bar rotating within the blender shell at a speed of 1500 -3000 rpm. The intensifier bar served both to shear the solids 30 and to spray the water into the charge in a finely divided form. The amount of water added was about 8% of the solids - 11 - l~'lO~73 weight and the time required for its addition was about 4 minutes. A further 1 to 2 minutes mixing time was allowed for the microagglomerates to achieve final size and compaction.
The product was then discharged onto a large tray, 5 spread out and oven dried at 80~C for 48 hours to ensure that drying was complete.
The dried product was then sieved to a size range of 125-500pm. A 100 g sample of the micro agglomerates was placed on a ceramic dish and heated for 25 minutes at 1260~C.
10 The sintered product was then subjected to several physical and chemical tests considered appropriate for determining its suitability as a feed material for reductive chlorination processing.
Visual inspection of the microagglomerates after 15 sintering revealed two obvious changes by comparison with the dried but unsintered material. Firstly, some shrinkage had occurred, either by a reduction in the internal voids of the microagglomerates or by a reduction of the intergranular voidage of the agglomerate mass on sintering. Secondly, the 20 colour of the material changed from a greyish brown to a reddish brown. Furthermore, the material assumed a glassy or reflective appearance in comparison to the dull surface of unfired material.
Microscopic examination of the sintered product 25 showed dense packing of the particles within the microagglomerates with abundant bridging between particles.
Electron microprobe analysis revealed no compositional differences between the material comprising the bridges and that of the particles. No appreciable degradation or 30 agglomerate-agglomerate adhesion was observed as a result of firing. X-ray diffraction analysis of the fired microagglomerates indicated major rutile and pseudobrookite phases, i.e., crystalline phases which could be formed from the original leucoxene alone.
The size of the product after firing was as shown in Table 3.

- 12 ~ 2 7 9 Size of Sintered Leucoxene Product from 75% -110 + 50 ,um and 25% -50 ,um feed agglomerated with 1% by weight Bentonite Binder and fired for 25 minutes at 5 1260~C.
SIZE (,um)CUMULATIVE % PASSING
500 100.0 355 97.4 250 78.7 180 27.4 125 0.0 A "strength text" was performed on the microagglomerates as follows: a microagglomerate was placed between two glass slides and weights were added until the microagglomerate first failed.
Failure first occurred at greater than 1 kg (i.e., approximately 10 N) for 300 ,um agglomerates. Fracture fragments were of similar size, i.e. there was little or no tendency to dusting. Calculations indicate that for the recorded strength it could be possible to store agglomerates without failure due to compressive forces in piles or storage binds of approximately 50 m in height.
A more quantitative and reproducible test for resistance to abrasion was determined by violently shaking one gram of a closely sized fraction of micro-agglomerates (-335 +250 ,um for 5 minutes in a cylin-drical tube 18 mm i.d. and 50 mm long with 3 ceramic balls 8 mm in diameter. During this test, the material was subjected to both impact and attrition. The average particle diameter after this test had reduced from 303 ,um to 170 ,um. This compares with the performance of a E~

- 12a - ~ 79 similar sample of the original leucoxene material which reduced to 220 ,um.
It may be concluded that the microagglom-erates represent an industrially useful material from the points of view of storage and transport.

_ _ . . . .. _, . . . .. _ .

Small samples (lOg) of microagglomerates were subjected to fluidised bed chlorination tests in a laboratory scale reactor at temperatures between 950 and 1100~C. The results showed that at greater than 50% completion of 5 chlorination:
(1) There was no indication of preferred attack on intergranular bonds. Rather the bonds appeared relatively more inert than the main mass of the individual mineral grains:
(2) Where the titania of the microagglomerates had been partially removed, an unreacted core of material of original appearance (apart from colour bleaching) remained within the microagglomerates. The pores of the affected outer shell were noticeably increased in size.
Table 4 provides initial and final size distributions for fired agglomerates which were taken to 89%
completion of chlorination in laboratory fluidised bed tests.
20 There is clearly little generation of -90~um material in chlorination, suggesting that high degrees of chlorination may be achieved without bond degradation or losses from reactors as fines carried in off gases. Similar results were obtained at up to 95% completion of chlorination.

. .. . . ... ... ..

- 14 - ~ 3 ~) 2~9 FLUIDISED BED CHLORINATION OF MICROAGGLOMERATES

Fluidising Conditions:
Temperature 1090~C
Time 70 min Chlorine Flowrate 1000 ml/min Calculated Gas velocity in bed 0.35 m/s Bed Height lOmm Feed Weight 10 g Bed Residue 0.9 g Carryover 0.25 g % Chlorinated 88.5 wt.~

SIZE RANGE FEED BED PRODUCT
(Jum) g % g %

+ 425 0.64 6.4 -425 +3553.24 32.4 - --355 +3003.25 32.5 0.10 11.1 -300 +2502.16 21.6 0.16 17.8 ~ -250 +1800.33 3.3 0.29 32.0 -180 +900.37 3.7 0.35 38.9 -90 0 . 01 0 .1 TOTAL 10.00 100 0.90 100 The fluidisation performance of the 25 microagglomerates was measured as a function of size and compared with the behaviour of theoretical spheres, petroleum coke and beach s4nd leucoxene. The result~, plotted a~
practical minimum fluidisation velocity in room temperature air against average particle diameter, are presented in - 15 - ~3.~279 Fig. 1. These results suqgest hiqher than expected minimum fluidisation velocities at smaller particle diameters and lower than expected minimum fluidisation velocities at larqer particle diameters. This behaviour may be explained partly by 5 size distribution effects and partly by density and surface shape and rouqhness effects. It suqqests that the chlorination process may be able to accept significantly larger agglomerate particles than is the case with conventional feeds, so affording the possibility of improved 10 process recoveries.

Approximately 10 kg of ground leucoxene were agglomerated and dried in the manner described in Example 1.
The microagglomerates were fed to a small pilot 15 scale fluidised bed furnace in which the bed temperature was maintained at a temperature of 1260~C. The operating parameters of the furnace were:
bed diameter 30 cm windbox temperature 1000~C
windbox fuel LPG
bed fuel coconut husk char superficial gas velocity 71 cm sec 1 in fluidised bed agglomerate feed rate 22 kg hr 1 In order to control both temperature and superficial qas velocity within the bed at the desired ranqe it was found necessary on this small equipment to enrich the inlet air with oxygen.
The average residence time of the material within 30 the bed was approximately 20 minutes.
The amount of bed material lost by entrainment in the off-gas was estimated at 4.5 wt.%. The size distributions of feed, product and carryover material were as shown in Figure 2.

~ lO2~

The product was subjected to the abrasion-attrition test described in Example 1. The result showed a reduction in average particle size from 303Jum to l90,um.

Agglomeration tests were carried out on a sample of rutile flour with the following size distribution;

Size Distribution of Rutile Flour Size (ym) Cumulative % Passing 96 98.8 64 88.6 48 80.2 32 58.6 24 42.0 12 27.1 : 8 22.3 6 18.4 4 16.0 3 12.4 2 10.5 1.5 9.9 9.1 Agglomeration was performed in an industrial ~ X~IX agglomerator, manufactured by Schugi Process Engineers of Lelystad, Netherlands at a solids feed rate of 840 kg per hour. Bentonite was premixed with the feed at 1%
addition and lignosulphonate was added as a 33 vol.% solution at 30 2.8 kg solids per hour. Moisture input in addition to lignosulphonate addition was 1 L min 1.

* l~ade mark . .

- 17 - 13~10~79 After continuous passage through the agglomerator and a fluid bed drying unit 67.5 wt.% of the product was in the size range +125 -500~m. Product coarser than 125~m diameter was collected for subsequent kiln based firing.
Firing of the agglomerates was conducted in a 3.6m long 0.23m internal diameter counter current oil fired rotary kiln. At a rotation speed of 2rpm and slope of one degree the agglomerate residence time in the 1260~C high temperature zone was approximately 20 minutes. A total of 60 kg of 10 agglomerates was fired in the kiln at a feed rate of 16.2 kg per hour.
Fine material in the feed and degraded material formed in firing were swept from the kiln by combustion gases, providing 69% recovery of feed in kiln products. Feed and~5 product particle size distributions are recorded below;

Size Distribution of Feed to and Product of Kiln Firing Size (~m) Cum % Retained Feed Fired Product 20850 9.07 6.67 600 19.65 16.31 425 32.20 30.86 300 46.85 50.62 212 67.11 82.82 25150 91.25 98.89 106 96.09 99.19 97.51 99.21 -75 100.00 100.00 Continuous agglomeration trials were performed in an 30 industrial blender manufactured by Patterson Kelley Pty. Ltd.
of Pennsylvania, U.S.A. The ground leucoxcene feed had the particle size distribution indicated below:

, . ~ . . . .

-- 18 - ~ ~ 4 0 2 79 Particle Size Distribution of Ground Leucoxene Size (~m) Cum % Passing 212 99.5 150 91.2 106 61.0 44.2 53 34.5 38 25.9 The blender was fed with ground leucoxene at 0.6 tonnes per hour with addition of bentonite at 6 kg per hour and organic binder (PVA) at 1.5 kg per hour. Moisture was added as 10% of feed weight via sprays mounted on the shaft of a set of high speed rotating blades within the agglomeration 15 chamber. Mineral residence time in the agglomerator was approximately 20 minutes.
The agglomerated product was dried in a tubular dryer to a maximum temperature of 80~C.
The particle size distribution of the dried product 20 is indicated below:

Size Distribution of Dried Agglomerates Size (~m) Cum % Passing 1000 100 . O
840 97.6 590 93.4 420 84.4 250 55.5 150 27.6 105 14.1 7.6 - 13~279 The dried agglomerated product was fed at 73 kg per hour to a 1250~C fluidised bed firing unit. The fluidised bed firing unit had a diameter of 0.46m and a height (above the distributor plate) of 0.56m. The fluidising gas was the air 5 rich combustion product of propane. Distillate was atomised into the base of the fluidised bed to provide additional heat by combustion with the oxygen remaining in the fluidising gases. Average residence time of the agglomerates in the fluidised bed was approximately 60 minutes.
Fine material present in the feed and generated in fluidised~bed firing was entrained in exiting combustion gases and removed via a hot cyclone. Only 17% of the feed reported in this "blowover" stream.
The particle size distributions of the fluidised bed 15 fired agglomerates and blowover are provided below:

Size Distribution of Products of Firing Size (~um) Cum % Retained Product Blowover 850 3.78 600 6.50 425 12.24 250 26.42 150 51.52 106 86.05 7.25 53 96.89 64.47

Claims

1. A process for increasing the particle size of fines of a titaniferous mineral containing more than 45% by weight titanium which process comprises:
mixing the fines with a binding agent and water to produce an agglomerate, drying the agglomerate and sintering the dried agglomerate at a temperature from 1000° to 1500°C to provide a sintered agglomerate.

2. A process according to claim 1, wherein the binder is capable of forming a glass or of exhibiting ceramic sintering properties when the agglomerate is sintered.

3. A process according to claim 1, wherein the binding agent is selected from the group consisting of:
1) colloidal silica, 2) silica, water soluble silicates or silica/fluorite mixtures, 3) bentonite, kaolinite and montmorillonite,

4) boehmite,

5) boehmite/silica mixture,

6) goethite,

7) lignosulphonate,

8) Sodium carbonate (saturated water solution),

9) Sodium silicate,

10) Group II element carbonate/clay mineral mixture,

11) molasses,

12) aluminum salt/organic amide mixtures,

13) titanium bearing organic and inorganic solutions,

14) polyvinyl acetate, and

15) water emulsified organic binders.

4. A process according to claim 1 or claim 2 wherein the binding agent is bentonite.

5. A process according to claim 1, 2 or 3, wherein the agglomerate is a microagglomerate formed by mixing the fines, binding agent and water by means of an impacting shearing action.

6. A process according to claim 1, 2 or 3, wherein the binding agent comprises from 0.5 to 5% by weight of the total weight of the fines and the binding agent on a dry weight basis.

7. A process according to claim 1, 2 or 3, wherein the water comprises from 5 to 15% by weight of the total weight of fines, binding agent and water.

8. A process according to claim 1, wherein the agglomerate is dried for less than 30 minutes at a temperature in the range from 75° to 150°C.

9. A process according to claim 1, wherein the agglomerates are sintered at a temperature in the range from 1200° to 1400°C for a period in the range from 5 minutes to 6 hours.

10. A process according to claim 1, wherein the titaniferous mineral comprises a mixture of fines and coarse particles.

11. A process according to claim 10, wherein the mineral is ground to produce fines.

12. A process according to claim 11, wherein the mineral is ground to particles having an average size in the range from 1 to 50 µm.

13. A process according to claim 1, wherein the sintered agglomerates have an average particle size in the range from 100 to 500 µm.

14. A process according to claim 13, wherein the agglomerates have an average particle size in the range from 150 to 250 µm.

15. A process according to claim 1, wherein the mineral is a detrital mineral.

16. A process according to claim 1 or 15, wherein the mineral contains more than 85% by weight of titanium dioxide.

17. A process according to claim 1 or 15, wherein the mineral is rutile, anatase or leucoxene.

18. An agglomerate produced by a process as defined in claim 1, 2, 3, 8 or 9.