CA1082880A - Process for the manufacture of aluminium chloride - Google Patents

Abstract

A B S T R A C T
In this process for the manufacture of aluminium chloride, aluminium oxide is transformed by dehydration into an activated form before being converted into the chloride by means of a chloridising and reducing gas mixture.

Description

-The invention concerns a process for the manufacture of aluminium chloride by the conversion of aluminium oxide using a mixture of chloridising and readucing gases. The term "aluminium oxide" will be used here to represent - all products which have been extracted from aluminium oxide bearing oresJ
without regard to the water content or the crystallographic structure. In the first stage of the process the aluminium oxide is brought, via dehydration, ; into a active form which is distinguished by its large true surface area and its small amount of residual water. This material of high reactivity is then employed in the subsequent chloridising stage.
The use of aluminium chloride for the industrial electrolytic production of aluminium failed up to now, because no reasonably priced aluminium chloride could be supplied in the required amounts (5 tonne of aluminium chloride is required for 1 tonne aluminium) and with sufficient purity. The incentive to find an industrial and economical method for the production of large quantities of aluminium chloride led to very many and varied ways of overcoming the problem being suggested. Up to now however, none of the proposals has achieved the desired aim.
In principle there can be considered to be two types of processes for chloridising aluminium oxide:
I Pure industrial grade aluminium oxide is coated or mixed with solid carbon and chloridised in a reducing reaction with a chloridising gas, ~ for example S2C12, COC12, NOCl, but above all with C12.
j The German patent DT-PS 1,237,955 describes a conversion process ' using chlorine, aluminium oxide and carbon of low (< 2.5 %) ash content, derived from oil, in a fluidized bed, at 450 to 650C. In order to ensure ` unimpeded progression of the reaction, without separation occurring it is ne-cessary to use aluminium oxide and oil-derived carbon of the same particle size.
The German patent DT-OS 2,244,041 describes a process, in which, in a first step at 750 to 1100C, a heavy oil is cracked in contact with an , ~

' ~

1~82880 aluminium oxide with a true surface aTea of at least 10 m2/g, until the par-ticles are coated with 15 to 24% car~on. In a second stage the carbon-coated aluminium oxide is chloridised in a fluidized bed at 450 to 800C.
II Pure industrial grade aluminium oxide is allowed to react with chloridised and reducing gases or mixtures of gases such as S2C12, C12 + S, , COC12, CL2 ~ CO or NOCl + CO.
The British patent GB-PS 718,773 describes the transfer of carbon monoxide and chlorine over a catalyst to form phosgene which then passes into - the chloridising unit.
The British patent GB-PS 668,620 works with aluminium oxide which is chloridised in a tunnel furnace on the addition of a melt of alkaline . aluminium chloride Cl part by weight of alkaline aluminium chloride to 1 part A12O3) which serves as a catalyst.
The German patent DT-PS 948,972 employs neither a catalyst nor phosg~ne, but requires the reaction to take place in a fluidized bed.
The German patent DT-AS 1,229,056 recommends the addition of 10 to 15 wt% of low ash carbon to the chloridising gases.
The German patent DT-PS 1,061,757 combines the fluidized bed with the alkaline aluminium chloride catalyst and requires special grinding up of the aluminium oxide. The particle distribution should effect a uniform distribution of the catalytic aluminium chloride within the three phase fluidiz-:-. ~
ed bed.
The use of pure industrial grade aluminium oxide for chloridising in accordance with the known processes of types I and II has the advantage that Nith the use of aluminium oxide bearing raw materials, prior or subsequent purification, is aYoided.
In the case of the processes in type I, carbon in the form of tar, pitch, asphalt, bituminous coal, or coke is used, whereby ~he porous products obtained from the coking should make the reaction with the chlorine-` 30 containing gas easier. The reaction is exothermic under 1000C. The coating

- 2 -. .

:. :
., .

1(~828BO

, of the aluminium oxide takes place either through mechanical mixing of carbon powder or particles and briquette making, or by the treatment of aluminium oxide with hydrocarbons in gaseous, solid or liquid form, whereby the hydro-i carbon is cracked and/or coked.
; Process of type I have the following disadvantages:
- They must be carried out in two stages.
- It is difficult to achieve a uniform deposit of carbon on the aluminium oxide.
- High demands are made with respect to the purity of the carbon;
it should be free of ash and sulphur, in order that the AlC13 formed does not have to be purified subsequently and only small losses of chlorine are suffer-ed.
In the processes of type II, aluminium oxide is converted using chloridising and reducing gases. The disadvantages mentioned in connection with the type I processes are for the most part avoided. However several new , problems are encountered:
- In general the chloridising process must be carried out at higher temperatures compared with carbon coated aluminium oxide particles, in fact approximately within 700 to 1000C.
- If one has to start with carbon monoxide, in theory only up to ` a half of the reducing effect of the carbon is made use of, or example at 600~C:
A12O3 + 3/2C + 3C12 = 2 AlC13 + 3/2 C02 ~H = -22 kcal/mol A1203 + 3C0 + 3C12 = 2 AlC13 + 3 C02 AH = -83 kcal/mol This need not be completely disadvantageous; as the enthalpy ~H
of the reaction in equation 2 shows, the reaction is still strongly exothermic.
It is also possible in many ways to make use Cafter appropriate scrubbing) of carbon monoxide, which occurs as a waste product in metallurgical processes, _ 3 _ ' 1082880 ~

e.g. in blast furnaces, as a reducing gas for the produc~ion of aluminium chloride.
W.D. Treadwell and L. Terebesi (Helv. Chim Acta 15, 1353, (1932)) already recognized, that the rate of reaction of aluminium oxide with chlori-; dising gases is dependent on the one hand on the heat treabm~nt conditions for the alumLnium oxide and, on the other hand, on the kind of gas employed. The ~?ration of treatment at 900 to looo& was varied between 2 and 10 hours.
Carbon monoxide and chlorine as well as phosgene were employed for the chlori-dation. In the case of the best heat treated aluminium oxide the conversion at 560C after 30 minutes amr~?nted bD 88% on using phosgene and 62% on using ~rbon nonoxide and chlorine. At higher temperatures, up to 1000C, there was cb~erved either only an insignificant increase in yield, or there was even an increase in the weight of the sample.
The aim of the invention presen o~? here is to develop a process for the production of aluminium chloride by converting aluminium oxide, in which process a complete and rapid conversion of the aluminium oxide is assured even at relatively low temperatures such as 350 to 400C, and for which one is not forced to provide carbon deposition, catalyst, phosgene or a special type or reaction unit. Most of the difficulties encountered in the chloridising steps of processes of types I and II should also be avoided.
~ This invention relates to a process for the manufacture of alumin--- ium chloride fram a ~rbon free hydrated aluminium oxide, comprising the steps of:
(a) activating said carbon free hydrated aluminium oxide by heat-ing fm m ambient temperature to a temperature of from 350& to 900C in one to 60 minutes to reduce the residual wa~r content to from 0.5 to 10 percent ", ~
by weight with referen oe to A1203 and to increase the surface area of the activated aluminium oxide to from 10 square meters to 450 square meters per gra~ Al203;

~ subsequently chlorinating in a catalyst free environment the ` activated carbon free aluminium oxide by bringing it to a temperature of from 350 C to about 800C and reacting it with a gaseous substance which acts both ~ ~J

' ~ :
.

~82880 as a reducing agent and as a chlorinating agent, while simultaneously main-t~;ning thermal eqyil;hrium during the reaction; and (c) condensing the gaseous reaction products to recover aluminium chloride.
At a giYen temperature and with a given reaction volume a higher through-put per unit time and space can be achieved wlth the prooess of the invention. }u:thermore the number of materials which can withstand the attack of the chloridising gases under the reducing conditions, is incomparably .

~ "
. , ., .

.. , s, ,...

;
.
. . .

.: ':
;~.' ~. ' ,' ~' ~
, ' ::
-4a-.~ï
::

.. . . .. , ~ . . .. ... .

,, : : - . : ' .

greater at lower temperatures.
At present aluminium oxide is normally produced by the Bayer process i.e. by extraction from bauxite using caustic soda solution followed by crystallization and calcination at approximately 1000C. The non-calcined product Al(OH)3, also call~d aluminium oxide in terms of the invention, is broken down in a first stage to A12Ox(OH)y. This highly active aluminium oxide has a very distorted crystal structure and exhibits a large true area. A pro-duction process and a comparable product are described for example in the German patent DE-OS 2,227,804.
Although the use of an iron hydroxide and low alkaline aluminium hydroxide (as is formed by settling the Bayer caustic solution) is particular-ly favourable, one can of course use all other kinds of industrial grade aluminium oxide with an alkali oxide content of 0 to 10 wt%, preferably 0 to 1%, expressed in Na20 with reference to Al(OH)3, and a particle size from 0.01 to 5 mm.
; The industrial grade aluminium hydroxide is broken down to an ~ active aluminium oxide in a dry or moist gas atmosphere, in particular in dry ,; air, by heating rapidly at various rates from room temperature to 350 to 900C, ;` in particular 500 to 800C, in 1 minute to 48 hours, preferably in 1 to 60 minutes, in particular in 1 to 15 minutes. There is a lower limit to the heat-ing up time used to bring the aluminuum oxide up to the desired temperature;
the upper limit to this heating up time is not of decisive l~portance. The '; active aluminium oxide has a true surface area of 10 to 450m2/g, preferably 't~'' 150 to 350 m2/g, and a residual water content of 0.5 to 10 wt%, with reference ;: to A12O , preferably 0.5 to 1 wt%, i.e. the active aluminium oxide must always ` have some, if small residual water content.
Instead of using a gas atmosphere with prescribed pressure, the ~ thermal transformation of aluminium hydroxide can also be carried out in vacuum.
The active aluminium oxide pxoduced by dehydration of aluminium -.
~. , : ~ ' :

hydroxide has a much distorted crystal structure and has a negligable amount of inactive aluminium oxide. The good, chemical reactivity is a prerequisite for t:he subsequent rapid chlorination.
I~e thermal transformation of aluminium hydroxide to an active aluminium oxide (A120x(OH)y) can take place in any suitable kind of reactor unitJ preferably however in fluidized bed, shaft or rota'ry furnaces.
The active aluminium oxide is chloridised either immediately after the removal of water by thermal treatment, or later. For this preferably one of the following chloridising gases or mixture of gases is used: chlorine and carbon monoxide, phosgene, nitrosylchloride and carbon monoxide, chlorine and sulphur, sulphur dichloride or mixtures of these which contain at least one of these chloridising and reducing components. If desired, low-ash carbon particles can be added to these gas mixtures. The chloridising is carried out at 350 to 800C, preferably 350 to 600C.
The mode of operation of the invention will now be explained using carbon monoxide and chlorine as an example representative for all gaseous atmospheres which can be used. The ratio of gases can be varied widely, from 90 mol% chlorine and 10 mol% carbon monoxide, up to 10 mol% chlorine and 90 ~ol% carbon monoxide, however a gas mixture containing 50 mol% chlorine and 50 mol% carbon monoxide is preferred. The reaction gases can be diluted with an inert carrier gas such as nitrogen, noble gases etc.
In the use of carbon monoxide and chlorine in the ratio 1:1, the ~-conversion of the highly active alumina described above is preferred, and in particular between 400 and 500C. The gas streaming down consists basically of aluminium chloride and oxides of carbon; chlorides of alkalis, iron, silicon and titanium, according to the purity of the aluminium oxide used, are present in only negligable amounts. The chloridising of active aluminium oxide can as with the thermal decomposition of aluminium hydroxide, be carried out in any suitable kind of reaction, preferably however in a fluidized bed shaft, or rotary furnace.

~. .
., .

- . -With respect to the residual water contentJ it must be emphasised that this does not concern adsorbed or absorbed water molecules but the hy-droxyl group built in to the crystal structure of the active aluminium oxide.
I all the water is driven off, for example by calcination at a higher tempera-tureJ then the active A120x(OH)y changes into ~-A1203 which is undesirable for the purposesof the process of the invention. On the other hand residual water given off during the chloridising stage leads to a loss of chlorine and carbon monoxide, in accordance with equation (3): -2 2 2 ~3) ~ GC500C) = -49.7 kcal/mol With 1 kg of water in the reaction going to completion 3.94 kg of C12 and 1.56 kg of CO are used up.
The rate of the chloridising reaction is a function of the true surface area of the aluminium oxide; this depends however on the amount of rèsidual water present. In the industrial application of the process of the invention therefore, reactor investments must be weighed against the costs in-curred by the losses in chlorine and carbon monoxide.
This invention will now be explained in greater detail with the aid of drawings and examples.

Figure 1: A schematic flow-chart of the industrial chloridising of aluminium oxide with chlorine and carbon monoxide in a series of fluidized bed reactors Figure 2: A schematic drawing showing a vertical section through the thermo balance used to study the chloridising process.
Figure 3: Curves showing the amounts of aluminium oxide conver-ted by chlorine and carbon monoxide at 400, 500, 600, 700 and 800C.
Figure 4: An Arrhenius plot of the rate of the chloridising ~, .
reaction.

: Figure 5: The dependence of the rate of reaction r on the partial pressure of CO ~PCl = constant) . ' .
- 7 ~

.

.

8Z88(~

Figure 6: The dependence of the rate of reaction r on the partial pressure of C12 ~PcO = constant~
Pigure 7: The dependence of the rate of reaction r on the ratio of partial preSsurespco PC12 Figure 8: Curves showing the amounts of aluminium oxide conver-ted at various partial pressure p ~ p Co C12 (PCo PC12 = 1:1) gas plpe-l1nes are indicated by single lines and pipes for the transport of solids with double lines in the f~ow-chart diagram in Figure l; the direction of flow is indicated by an arrow.
The fluidized bed reactor 10 is continuously fed from a silo 11 with reduction plant grade aluminium hydroxide. This hydroxide is heated from room temperature to about 400C by hot air from the reactor 13 and partly dehydrated. The hot aluminium oxide, with a residual water content of 5 to 10 wt% ~hen enters the fluidized bed reactor 13, where at a temperature between 400 and 800C it is turned into the final active form A120x(OH)y and then transported into the fluidized bed reactor 14. There the active aluminium oxide is sufficiently cooled by air at room temperature, that in the subsequent fluidized bed reactor 16 the chloridising reaction proceeds at thermal equili-brium, i.e. such that it need be neither cooled nor heated. The preheated dry air from reactor 14 is led together with the fresh air 17 to the air heating unit 18, from there into the reactor 13 and then into the reactor 10 where it is expelled as moist air 12.
In reactor 16 the actual chloridising takes place between 350 and 800C. The gas coming out consists mainly of aluminium chloride and oxides of carbon, as well as solid particles carried away by the gas stream and liquid droplets containing the elements aluminium, sodium, oxygen and/or chlorine;
the gas is led through the cooler 19 and enters the separator 20 at a tempera-ture above the condensation point of aluminium chloride (183C). In this the liquified and solidified components are separated and partly recycled by a distribution 21 via the reactor 16. The aluminium chloride is condensed in the ' condensor 22; the solid pxoduct of the react;on is co-llected in a container 23, whilst the remainder of the gaseous components pass on to a waste gas scrubbing unit 24, and part recycled with fresh chlorine and carbon monoxide 25 ih the reactor 16.
Example 1 In the course of the chloridising reaction the loss of weight of the highly active aluminium oxide was measured by a thermo-balance as a func-tion of the reaction time, and its specific surface area determined before and during chloridising by the BET method.
: 10 The iron and low alkali aluminium hydroxide, produced by the Bayer , process and used for the thermal decomposition and chloridising, has an aver-age particle size of 80 ~m. An analysis of this product gives the following result:
`' Moisture (105C) 0.251%
. SiO2 0.019%

2 3 0.007%
2 0.003%
P2O5 0.001%

~25 0.001%
CaO 0.012%
Na20 ~total) 0.410%
density 2.42 g/cm3 The weight loss experienced by the aluminium oxide during the !. , chloridising was measured in the low temperature quartz furnace 26 shown in - Figure 2. 145 to 150 mg of aluminium hydroxide were weighed out in the quartz `; crucible. ~or the thermal decomposition, a stream of nitrogen gas at a rate of 16 l/h was introduced into the quartz furnace through the inert gas inlet 28, whilst through the inlet 29 for reactant gas a nitrogen gas stream of 4 l/h was introduced, and the total amount of gas escaped through the outlet 30 in the side wall of the quartz furnace. The heating coil 31 was switched on .,~ .

and the furnace heated up within 30 minutes to a constant temperature for reaction, between 400 and 800C. The actual heating up time is only 10 to 15 minutes; the 30 minutes are required to achieve a constant temperature for reaction within + 4C. The thermal decomposition of aluminium hydroxide to an active aluminium oxide begins at 250C. After 30 minutes the stream of nitro-gen through the inlet 29 for the reactive gas was turned off and chlorine and carbon monoxide in a molar ratio of 1:1 (each 2 l/h) were introduced through this inlet. At the start of the chloridising reaction the temperature rises slightly as a result of the reaction being exothermic. The aluminium chloride produced sublimates from the hot reaction zone and condenses as a white sub-limate on the cold gas outlet 30 in the quartz inner tube 32 and on the water cooled furnace outlet 33 which is anchored to the base 34 of the balance. -The measurements were made around 400, 500, 600, 700 and 800C.
In further trials at the same temperatures the true surface area of the active aluminium oxide, with and without the chloridising reaction was measured by the BET method as a function of the reaction time and decomposition time. -The following were carried out for this purpose:
a) Aluminium hydroxide was decomposed to highly active A120x(OH)y at temperatures between 400 and 800C, using the same rate for heating up as with the samples for the chloridising trial, and the true surface area measured after various reaction times at 400 to 800C.
b) Aluminium hydroxide was converted to active aluminium oxide as in the above trial and chloridised for various intervals of time between 400 and 800C. The true surface area of the samples obtained was likewise deter-mined.
The reduction in the true surface area of the actl~e A120xCOH)y and the amount of aluminium oxide converted during the time of the reaction or conversion are shown in Figure 3.
The curves 41 (400C), 51 (500C), 61 C600C), 71 C700C) and 81 ,~ . . . .

108288~

(800C) show the change in true surface area of the active aluminium oxide as a function of the time at 400, 500, 600, 700 and 800C. The structural change takes place only by heat treatment, no chemical reaction takes place.
In the curves 42 C400C), 52 (500C), 62 (600C), 72 (700C) and 82 ~800C) the reduction in true surface area of the active aluminium oxide is shown as a function of the reaction time during the chloridising.
The curves 43 (400C), 53 (500C), 63 (600C), 73 (700C) and 83 (800C) show the amount of aluminium oxide converted as a function of the ~ reaction time on chloridising with C12 and CO, whereby the curves 63, 73 and ;~ 10 83 lie within the error limits.
It can be concluded from Figure 3 that the true surface area of A12Ox(OH~y is hardly dependent on the further treatment after the heating up period and in the case of a purely thermal reaction remains practically unchan- -ged during the whole reaction time.
During the chloridising stage the true surface area of the ;~ .
A12Ox~OH)y falls markedly.
The reaction rates obtained from the experimental data in Figure

3 are shown in Figure 1.
In all trials between 500 and 800C the influence of gas film diffusion was observed at the beginning of the reaction; at 400C the gas film diffusion for a conversion of 2 wt%/min., with reference to the weighed amount which lay between 145 and 150 mg, does not yet limit the rate of reaction. In the trials between 500 and 800C, towards the end of the reaction, with a ; conversion of approximately 1.5 wt%/min., with reference to the starting amount between 145 and 150 mg, the gas film diffusion does not limit the rate of the chemical reaction with respect to the true surface area.
All the reaction rates obtained in example 1 were calculated from the conversion curves at a conversion of approximately 1.5 wt%/min., with reference to the initial amount which was between 145 and 150 mg. From this it can be seen that the calculated amounts, converted between 500 and 800C, are . . s 108Z8~

.
too low. The reaction actually goes to completion in a much shorter time.
Therefore, for example, under exactly the same above reaction conditions at 700(, a sample of 10 mg active A120 COH)~ is 98.9% converted in 18 minutes.
Pigure 4 shows the Arrhenius plot of the logarith~ of the reaction rate r versus the reciprocal of the absolute reaction temperature. The activation energy of the chloridising reaction can be determined from the slope.
The change in the activation energy at 600C allows one to conclude that the active oxide transforms to a more stabile modification at higher temperatures.
The reduced activity of aluminium oxide transformed at 800~C and chloridised , . , - 10 at 400C indicates the same.

Table 1 ., ::
Trial No. Temp. of True Residual Chlori- Rate of conver- Surface Water dising reaction . sion area tempera-~C) ~m2/g) (%) (C) (g/(m2.min)) Example 1 1 400 280 7.4 400 0.78 . 10-4 2 502 281 2.4 502 2.4 . 10-4 3 604 258 1.6 604 5.5 . 10-4

4 700 215 0.9 700 5.8 . 10-4 803 178 0.9 803 8.8 . 10-4 Example 2 6 803 187 0.6 400 0.42 . 10-4 Example 2 The starting material, amount of sample and apparatus were the same ;;-as in example 1. 145 mg of aluminium hydroxide were heated in a 4 l/h stream of nitrogen from room temperature to 800C within 15 minutes, whereby a ther-mal conversion took place to give active A120x(OH)y which was then chlori-dised at 400C in a gas mixture of chlorine and carbon monoxide each at 2 l/h.
The reaction rate entered in table 1 was obtained from the amount converted . .

, .
.~ . - , . . . .

~08288~

in the first minute of the reaction and the relevant true surface area.
- Converted after l minute: 0.79 wt%
- True Surface area 180 m2/g This example shows that the rate of the chloridising is a function of the true surface area.
Table 1 shows that a good solution to the problem of chloridising aluminium oxide with C12 and C0 is first to transfo~m the aluminium hydroxide at 800C into an active A120 COH)y and then to chloridise this at 400 to 500C.
Under these conditions the true surface area is large (187 m2/g) and the residual water content, which leads to a loss of chlorine and carbon mono-, side, is small ~0.6 wt%). The chloridising reaction takes place quickly enough at low temperatures, at which material problems are easy to solve~
Example 3 The starting material, sample size and apparatus were the same as i in example 1. 145 mg of aluminium hydroxide were heated within 15 minutes from room temperature to 800C in a 4 l/h stream of nitrogen gas, whereby an active A120x(O~)y resulted from the thermal decomposition of the sample. The quartz furnace was allowed to cool to the constant reaction temperature of 400C within 30 minutes and the active aluminium oxide chloridised by passing a mixture of chlorine, carbon monoxide and nitrogen over the sample.
The dependence of the reaction rate on the partial pressure used was investigated for a mixture of carbon monoxide and chlorine at 400C. The desired partial pressures were arrived at by the provision of the app~0priate flow rates ~total 51/h) of the individual components. Nitrogen was used for dilution purposes. Standardized rotameters were used to measure the flow rates.
Under the experimental conditions given in table 2 the conversion of active aluminium oxide at 400C is not limited by gas film diffusion. The rates of reaction are calculated from measurements of the amount converted and - 30 the true surface area of the aluminium oxide at the start of the reaction, and after 15 minutes and 30 minutes operation time, and the average of these -three values is taken. The results of these chloridising trials are given in table 2.
T ble 2 ,, .

Trial PC12 Pco pN2Rate of Reaction (atm) (atm) (atm)Cg/m2. min) 1 0.472 O.O9S 0.3801.20 . 10-5 2 0.471 0.191 0.2851.95 . 10-5 3 0.472 0.286 0.1892.80 . 10-5 4 0.472 0.3~0 0.0953.44 . 10-5 ~-0.471 0.476 0 4.53 . 10-5 6 0.376 0.476 0.0954.10 . 10-5 7 0.285 0.475 0.1873.54 . 10-5 8 0.189 0.474 0.2843.18 . 10-5 9 0.095 0.474 0.3781.97 . 10-5 .
The general form of the rate equation is:

r = _ 1 dG = k P C0 P C12 where, 2 r: rate of reaction (g/m . min) s: true surface area of A120x~OH)y(m2/g) - m,n: order of the reaction dG : weight change in A120x(OH)y per unit time (glmin) k: overall reaction rate constant (g/m2.min.atmCm n)) , . . 1 .
Pco: partial pressure of carbon monoxide (atm) PCl : partial pressure of chlorine (atm) `` The results of table 2 are summarized in figures 5 and 6. They show the dependence of the rate of ~he reaction, of aluminium oxide with ;' ~' ': . ,' 1~8Z880 ~:

, .
.' chlorine and carbon monoxide at 400C, on the partial pressure. In a double , logarithmic plot straight lines are obtained, the slopes of which give the orders m and n,of the reaction. The value of m taken from Figure 5 is 0.807 and for n from Figure 6 is 0.504:
The reaction, rate constant k was calculated from the reaction rates of the trials 1 to 9 by substituting for P CO and pncl in the rate equation.
The following empirical equation describes the influence of the ' partial pressure of chlorine and carbon monoxide on the conversion rate of ~,; 10 aluminium oxide decomposed at 800C and chloridised at 400C:

''' r = 1.18 . 10-4 . pco807 . Pci504 ~5) ', In connection with the reaction rate equation it must be mention-ed that in the trials 1 to 9 the reaction products aluminium chloride and carbon dioxide have been immediately removed in the excess reaction gas pass-ing over the A12Ox(OH)y. Tentative trials however allow one to suppose that the presence of AlC13 accelerates the conversion of A12O3.
' The reaction rate r for the conversion of active aluminium oxide ,' as a function of the ratio of the partial pressures of carbon monoxide and ,' chlorine is shown in Figure 7. From this it can be seen that the chloridising of aluminium oxide is preferably carried out with a gas mixture of 50 mol%
chlorine and 50 mol% carbon monoxide. Any dilution of this gas mixture with ~, an inert gas and/or a reduction of the chlorine or carbon monoxide partial .
'~" pressure leads to an obvious slowing up of the reaction.

Example 4 The starting material, sample size and apparatus were the same , as in example 1. 1.45 mg of aluminuum hydroxide were heated in a 4 l/h stream ;'' of nitrogen from room temperature to 600 within 30 minutes~ whereby an active A12Ox(OH)y was obtained from the thermal decomposition of the sample and was '. subsequently chloridised.

"` 30 Figure 8 shows the curves obtained by plotting the weight % of ''' 1~8Z880 ,. :
converted Al203 as a function of the reaction time, in two chloridising trials with a carbon monoxide to chlorine ratio of l and at various partial pres-sureS of Pco + PC12 Curve 63 shows the conversion of A120x(OH)y at 600C using a gas mixture of 50 mol% chlorine and 50 mol% carbon monoxide at a total pres-sure of 0.947 atm as a function of the reaction time (compare table 1, example l, trial No. 3).
Curve 67 shows the conversion of active A120x(OH)y at 600C using :
a gas mixture of 5 mol% chlorine, 5 mol% carbon monoxide and 90 mol% nitrogen at a total pressure of 0.947 atm as a function of the reaction time.
Details of the experimental conditions and the results of these chloridising trials at 600C are summarized in table 3.
Table 3 Gas composition Ptot Pco PCl2 2 (atm~ (atm)(g/m . min) ., _ , 0.947 ~ 0.947 ~ .

5% Cl2 0.0947 0.0947 1.67 . lO
.,,, .' :
The reaction rate equations (5), Eigures 8 and table 3 show, that in the region of the check on the chemical reaction rate, an increase in the partial pressures of the reactive gases allows the chemical reaction rate r :: to rise markedly. It follows from this, that elevated pressures, up to 40 atm, in particular at 1 to 10 atm, are preferred.

, ''' .
J

~' ,

Claims (24)

THE EMBODIMENTS OF THE INVENTION IN WHICH AN EXCLUSIVE
PROPERTY OR PRIVILEGE IS CLAIMED ARE DEFINED AS FOLLOWS:

1. A process for the manufacture of aluminium chloride from a carbon free hydrated aluminium oxide, comprising the steps of:
(a) activating said carbon free hydrated aluminium oxide by heating from ambient temperature to a temperature of from 350°C to 900°C in one to minutes to reduce the residual water content to from 0.5 to 10 percent by weight with reference to Al2O3 and to increase the surface area of the activated aluminium oxide to from 10 square meters to 450 square meters per gram Al2O3;
(b) subsequently chlorinating in a catalyst free environment the activated carbon free aluminium oxide by bringing it to a temperature of from 350°C to about 800°C and reacting it with a gaseous substance which acts both as a reducing agent and as a chlorinating agent, while simultaneously maintaining thermal equilibrium during the reaction; and (c) condensing the gaseous reaction products to recover aluminium chloride.

2. A process according to claim 1 wherein the aluminium oxide comprises industrial grade aluminium hydroxide which is dehydrated and chloridized.

3. A process according to claim 2 in which industrial grade aluminium hydroxide with an alkali oxide content of0 to 10 wt%, expressed in Na2O with reference to Al(OH)3 is used.

4. A process according to claim 3 in which industrial grade aluminium hydroxide with an alkali oxide content of 0 to 1 wt% is used.

5. A process according to claim 1 in which aluminium oxide with a particle size of 0.01 to 5 mm is used.

6. A process according to claim 5 in which the aluminium oxide is dehydrated by heating from room temperature to a temperature between 350 and 900°C in vacuum.

7. A process according to claim 6 in which the dehydration takes place in dry air.

8. A process according to claim 1 in which the dehydration takes place by heating from room temperature to 500 to 800°C.

9. A process according to claim 1 in which the heating up period is between 1 and 60 minutes.

10. A process according to claim 9 in which the heating up period is between 1 and 15 minutes.

11. A process according to claim 1 in which dehydration is carried out until the residual water content is 0.5 to 1 wt%, referred to Al2O3.

12. A process according to claim 11 in which dehydration is carried out until a true surface area of 150 to 350 m2/g Al2O3 is obtained.

13. A process according to claim 1 in which the dehydration is carried out in a fluidized bed furnace, shaft furnace or a rotary furnace.

14. A process according to claim 1 in which dehydrated aluminium oxide is converted using one of the gases, or gas mixtures, chlorine, and carbon monoxide, phosgene, nitrosylchloride, carbon monoxide, chlorine and sulfur or sulfur dichloride.

15. A process according to claim 14 in which a gas mixture from 10 to 90 mol% of chlorine and 10 to 90 mol% carbon monoxide is used.

16. A process according to claim 15 in which a gas mixture of 50 mol%
chlorine and 50 mol% carbon monoxide is used.

17. A process according to claim 14 in which the chloridising and reducing gas is diluted with an inert gas.

18. A process according to claim 17 in which nitrogen or a noble gas is used as the inert gas.

19. A process according to claim 14 in which the reaction gases are used at a sum of partial pressures of 0.1 to 40 atm.

20. A process according to claim 19 in which the reaction gases are used at a sum of partial pressures of 1 to 10 atm.

21. A process according to claim 20 in which particles of low ash carbon are used with the chloridising and reducing gas.

22. A process according to claim 15 in which AlCl3 is added to the gas mixture.

23. A process according to claim 22 in which the chloridising is carried out at 350 to 600°C.

24. A process according to claim 14 in which the chloridising is carried out in a fluidized bed reactor, solid bed, or rotary reactor.