Note: Descriptions are shown in the official language in which they were submitted.
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
2 7 3
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.
_. .
7 3
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
EXAMPLE 1
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.
-- 10 --
~ ~0~73
TABLE 1
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
TABLE 2
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
TABLE 3
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.
_ _ . . . .. _, . . . .. _ .
- 13 - 1340273
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
TABLE 4
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.
EXAMPLE 2
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~
- 16 - -
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.
EXAMPLE 3
Agglomeration tests were carried out on a sample of
rutile flour with the following size distribution;
TABLE 5
Size Distribution of Rutile Flour
Size (ym) Cumulative % Passing
128 100
96 98.8
64 88.6
48 80.2
32 58.6
24 42.0
16
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;
TABLE 6
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
TABLE 7
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:
TABLE 8
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
-- 19 --
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:
TABLE 9
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
