GB1576531A - Recovery of tin from ores or other material - Google Patents

Description

(54) RECOVERY OF TIN FROM ORES OR OTHER MATERIALS (71) We, COMMONWEALTH SCIENTIFIC AND INDUSTRIAL RESEARCH ORGANIZATION, a body corporate established under the Science and Industry Research Act 1949, of Limestone Avenue, Campbell, Australian Capital Territory, Commonwealth of Australia, and ABMINCO N.L. a company incorporated under the laws of the State of Victoria, of 233 Collins Street, Melbourne, in the State of Victoria, Commonwealth of Australia, do hereby declare the invention, for which we pray that a patent may be granted to us, and the method by which it is to be performed, to be particularly described in and by the following statement: This invention relates to an improved process for recovering tin or other non-ferrous metal values from low-grade ores, concentrates, or mineral mixes which consist mainly of iron sulphide minerals.

Conventional tin smelting processes are applied to essentially sulphide-free, high grade concentrates such as those derived from alluvial sources, containing tin in excess of 60%.

These concentrates, which are prepared by conventional mineral dressing methods are treated in a two stage reverberatory furnace operation. The concentrates are charged to the reverberatory furnace at 12000C with fluxing materials to cause the charge to melt and form a fluid slag which is reduced by a suitable reductant such as coal to produce crude tin. The tin product from the first stage is sufficiently low in iron and other impurities to permit refining by conventional methods. The slag which contains typically 5 to 10% tin, must be granulated and treated in the second stage to recover this tin.It is mixed with additional reductant and treated at approx. 1400"C, to produce a discard slag containing 1.5 to 3 percent tin and a metallic product with a high iron content (hard-head) which is recycled to the first stage.

Low grade concentrates from hard-rock sources not only contain larger quantities of gangue minerals (and therefore produce large volumes of slag per unit of crude tin produced) but usually have higher iron-to-tin ratios, which leads to the formation of large quantities of high melting point hard-head with high iron content which cannot be dealt with in the normal two stage recycle. These two factors make it uneconomical to treat low grade concentrates in conventional smelters.

In the treatment of low grade concentrates it is usual to employ processes which involve a slag fuming operation to remove tin from the system either as a fume or stannous oxide SnO (by appropriate control of oxidation/ reduction conditions) or as a fume of stannous sulphide SnS (by injection of sulphur or pyrites to control sulphidizing conditions, together with coal or a hydrocarbon fuel to control reduction conditions). Careful control of the charge composition and rates of addition of reagents is required to maintain fuming conditions without matte formation. Any matte which is formed is treated separately to recover its tin values or recycled as a sulphidizing agent. This type of slag-fuming operation is generally employed as a slag cleaning stage for slags derived from the first stage of conventional smelting operations.

It allows the iron entering the smelter to be discharged in the discard slag, and thus avoids the buildup and excessive recycle of iron. In all such slag fuming processes the fuel and sulphidizing agent are injected into the slag layer, below its surface, and matte formation is avoided where possible. It is possible, though not usual, to charge concentrates which may or may not contain iron sulphides, to the fuming furnace.

This type of process is represented in the prior art by: 1. United States Patent 2,304,197 Dec. 8, 1942, W.H. Osborn.

2. D.V. Belyayev, The Metallurgy of Tin, Pergamon Press, Oxford, 1963, p. 88.

An alternative approach dispenses with both the preconcentration step and the conventional smelting step; it is basically an ore fuming process which is applicable to ores or blends which consist mainly of iron sulphide minerals with relatively small quantities of other minerals, including cassiterite.

It will be appreciated that the avoidance of the preconcentration stage and the use of a high-temperature fuming process on the whole of the ore as mined introduces some very special economic restraints. For a given production of tin from a 1 % tin ore, as distinct from say a 20% concentrate, the total material throughput will be twenty times greater. Thus capital costs must be optimized and the specific throughput of the plant must be as great as possible, to retain economic viability.

Of even greater importance from the economic point of view is the cost of the energy required to heat the whole of the ore to the fuming temperature. In an ideal ore fuming process this heat must be supplied by burning of the components of the ore without the need for additional fuel. These processes are in general therefore only applicable to ores containing sufficient iron sulphides to provide all the process heat from their combustion. (Of course, if the ore is supplemented by concentrates containing a greater proportion of tin, it may become economical to use some supplementary fuel).

In brief, an ore fuming process should be a high intensity process with high specific throughput, and should be autogenous.

Two distinct variants of sulphide ore fuming processes have been described in the prior art.

These are: (1) Non-slagging processes which are applied to iron sulphide ores of the type for which the present invention is intended. In these processes the essential feature is that the temperature is maintained at a sufficiently high level to permit volatilization of tin sulphides to occur, but at a sufficiently low level to avoid incipient melting of the charge. The processes may thus be carried out in fluidized beds, multihearth roasters or shaft furnaces and the residues are free-flowing particulate solids. Such processes are represented in the prior art by the following patents: 1. United States Patent 2,600,351 - Wells, Thompson & Roberts, Dorr Company 1952 2. United States Patent 1,847,991 - Sulman & Picard 1932.

3. Australian Patent 3735/26 - Krupp 1926.

(2) Slag forming processes applied to sulphide ores. This group of processes like group 1 is also applied to ores consisting mainly of iron sulphide minerals. In this group the temperature of the charge is allowed to rise sufficiently due to burning of the iron sulphide minerals that melting and slag formation occur and the final residue is an iron silicate slag. Formation of a matte is avoided if possible, but if it does form, it is collected in the hearth, tapped from time to time and treated to recover its tin content. This type of process is represented in the prior art by Trostler and Carlsson, United States Patent 2,219,411 (1940) and Australian Patent 109,112 (1939).

The present invention falls into this group, and is related to the process of Carlsson and Trostler, in that it is intended for the treatment of ores or concentrates consisting largely of iron sulphides with sufficient silica naturally occurring or deliberately added to cause all the iron produced in the residue from burning of the iron sulphides to form an iron silicate (fayalite) slag. The molten slag is the residue from the process and is tapped for discard.

In the treatment of certain sulphidic tin ores from the West Coast of Tasmania, conventional mineral dressing treatments failed to achieve recoveries better than 30% with a concentrate grade of 30 percent tin. Thus conventional treatment involving preconcentration and smelting (even those smelting processes applicable to low grade concentrates) could not be applied.

Our early attempts towards development of an autogenous ore fuming process for treat ment of these pyritic tin ores, involved the non-slagging approach, i.e. heating the ore in such a way that the tin was sulphidized and volatilized without allowing the temperature to rise to the level where slagging reactions could cause agglomeration and sticking of the charge. In this process, the volatilization and recovery of tin is achieved in the following way: (1) The ore is fed to a shaft or multi-hearth furnace co-currently with hot gases derived from combustion of residues in later stage (3). Co-current operation retains the partial pressure of sulphur released from the pyrite as the temperature of the ore rises.

(2) The gases from the co-current volatilization (stage (1)) containing sulphur, sulphur dioxide, carbonyl sulphide (COS), nitrogen, carbon dioxide and stannous sulphide, are collected and burnt to convert all sulphur and SnS to SO2 and SnO2. The gases are cooled and the SnO2 fume collected.

(3) The heated residue from stage (1) which contains pyrrhotite, silica and minor amounts of other minerals is passed through a suitable lock system into a reactor where it is burned (by counter-current contact in a separate compartment) with air diluted by cooled tail gases from stage (2). The proportion of diluent is adjusted to limit the temperature of the burning mass to below the temperature of incipient fusion, while producing a sufficient volume of combustion gases of sufficiently high temperature to carry out the co-current heating of the ore.

When attempts were made to conduct this process autogenously in a simulated practical reactor system it was found to be impossible to achieve a suitable combination of gas temperature and gas volume from stage (3) without causing slagging reactions and consequent sticking and agglomeration of the reacting mass. Furthermore, even in the absence of these problems the practical difficulties associated with (a) balancing of reaction rates in the two stages, (b) transfer of gases at very high temperatures from stage (3) to stage (1) and pressure balancing between reaction zones, and (c) collecting fume from excessively large volumes of recycled gas, rendered the process unattractive.

The blast furnace pyritic smelting process described by Carlsson and Trostler appeared to offer the only remaining possibility in the prior art for the treatment of these ores. However tests conducted in a small shaft reactor showed that incipient fusion above the melting zone caused bridging of the charge, and the process could not be made to operate. There is no evidence in the literature that this blast furnace process was ever operated commercially.

The problem accordingly remained of finding a process for concentrating cassiterite from the ore which is not subject to the disadvantages of the prior art process discussed above.

It has now been surprisingly found that the problems of heat and mass transfer described above can be overcome in a process in which the ore is added to a refractory-lined reactor containing a, pool of liquid iron sulphide matte overlain by a liquid iron silicate slag, and the tin sulphide volatilization is achieved by a process akin to matte conversion. Such a process may be referred to as "matte fuming".

Thus, the present invention provides a process for extracting tin values from a tin bearing iron sulphide ore or mineral-mix, comprising charging said ore or mix in particulate form into a refractory-lined reactor containing a pool of molten matte so that said iron sulphides in said ore or mix become molten and enter the matte pool, blowing air or other oxidizing gas through the matte pool to oxidize said iron sulphides thus producing iron oxide and generating the heat required to smelt the charge and drive off the tin values from the charge in the form of volatile sulphides and oxides, said iron oxide combining with silica added in the charge to form a layer of molten slag which floats on the matte bath, the conditions in the reacting mass being such that the oxidizing gas blown into the reactor comes first into contact with the pool of molten matte, thus stabilizing the oxygen and sulphur pressures of the gases, after contacting the matte, in a suitable range to sulphidize and volatilize the tin in the charge without the need to add a carbonaceous reductant.

The iron oxide produced by the combustion of the matte combines at the surface of the matte with silica from the charge to produce an iron silicate slag which is maintained saturated with silica. This slag is the final residue from the process, and is discarded when its tin content has been lowered sufficiently by the fuming reactions.

The large quantity of gas which passes through the matte to bring about its oxidation sweeps out or scavenges the tin sulphide, which collects in the matte, by maintaining a very low partial pressure of tin sulphide in the large volume of gas bubbles. Our observations have shown that a substantial fraction of the tin contained in the particles of charge can be sulphidized and volatilized while the particles are actually being heated up in the reactor to the temperature of the bath and before they are incorporated into the slag or matte phases.

Thus, although published studies of the vapour pressure of tin sulphides over tin-iron mattes would lead to the expectation of a substantial build up of tin in the matte and a corresponding equilibrium partition of tin in the slag, it is possible by correct operation of the present process and by correct sizing of charge particles to limit the extent to which the tin enters the matte.

However, this does not set a limit on the production rate since the tin which does enter the matte is removed during the conversion of the matte. Furthermore, the gas has its oxygen and sulphur partial pressures fixed by the fact that it has equilibrated with the matte which is in turn equilibrated with the silica-saturated slag. Partial pressure for oxygen is approximately 10-s atm and that for sulphur 10.1.5 atm and, therefore, very efficient fuming of tin from the slag is promoted as the large volume of gas passes through it, maintaining a very low tin concentration in the slag.

Some of the reactions involved in the process are as follows.

Heating of the charge of ore or concentrate: FeS2 = FeS + x) + (1 - x)S (1) FeCO3 = "Fev" + C02 (2) SnO2 + S2 = SnS + SO2 3 2FeO + 1-S2 = 2"FeS" + SO2 4 Oxidation and slagging reactions: "FeS" + 1102 = "FeO" + SO2 (5) "FeO'-' + SiO2 = 2FeO. SiO2 (6) 2"FeS" + SiO2 + 302 = 2FeO. SiO2 + 2SO2 (7) C + O2 = CO2 8 2FeS + 3402 = Fe2O3 + 2SO2 (9) Reaction (7) is a combination of reactions (5) and (6) and is written for simplicity in stoichiometric form.However, experimental observations of a number of workers have established that some labile sulphur will always be evolved principally due to the nonstoichiometry of the slag formed.

In the operation of the actual process the matte and slag phases are much more complex than simple ferrous sulphide and ferrous silicate. Both are in fact "oxysulphide" phases containing ferrous oxide, ferric oxide, silicon and sulphur. The oxidation potentials of the matte and slag are related to their ferric oxide activities.

Our experiments have shown that the oxidation of some of the iron to the ferric state results in an enhancement of the heat released by the process, and it is advantageous to drive the process at a high rate to optimize the degree of oxidation of iron to the ferric state. The limit to the permissible extent of oxidation is set by the requirement to maintain phase separation between the matte and slag phases.

A typical process according to the invention may be carried out as follows: Stage (1): A bath of iron sulphide matte saturated with ferrous oxide is maintained at a temperature of 12500C in a refractory-lined vessel equipped with lances or tuyeres to permit air to be blown through the matte. At the start of a new batch this matte, overlain by a thin layer of slag, remains from the previous batch. Its iron oxide content is fixed by equilibration with a silica saturated slag. Air is blown through the matte via the lances or tuyeres to convert iron sulphide to iron oxide which enters the slag.

Stage (2): Pyritic tin ore or a blended mixture of tin-bearing materials containing appropriate amounts of silica and iron sulphides (hereinafter referred to for convenience as "ore") is crushed to say -5mm and added slowly to the reactor while maintaining the air blast to generate heat by combustion of the matte. In the preferred method of operation the ore is screened to separate the - 1 mm fraction which is injected directly into the matte layer through the lances or tuyeres along with the air and supplementary fuel as required. The + 1 mm fraction is charged onto the surface of the slag by means of an appropriate feeder. The rate of charge addition is adjusted to match the rate of production of heat and so maintain the matte and slag in the molten condition.The iron sulphide in the charge enters the matte, and during heating to the melting temperature, labile sulphur in the pyrite fraction of the charge is released. This is partly evolved as elemental sulphur and partly consumed in sulphidizing the iron oxides and carbonates and tin oxide in the ore particles and in the matte and slag baths.

The net result is a large increase in the amount of matte consequent upon the addition of the charge. The silica fraction of the ore particles combines with the iron oxide which is produced by conversion of the matte, forming more slag which gradually increases in volume at the expense of the matte. If the charge is not self-fluxing, additions of pyrrhotite (which may contain tin) or siliceous materials may be made to adjust the iron-to-silica ratio to produce a suitable slag. The violent agitation of the bath causes the matte and slag to splash over the added ore particles and provides rapid heat transfer.

Stage (3): After completion of ore addition, blowing is continued until the stoichiometric oxygen has been supplied to convert all the new iron sulphide matte resulting from the addition of the charge of ore. During this stage the tin which enters the slag and matte is fumed off.

Stage (4): -The bulk of the slag is tapped for discard leaving the original quantity of matte overlain by some slag for the next batch.

During all the above stages the air blast may be continued to maintain a continuous SO2 feed to an acid or sulphur plant.

The gases leaving the bath contain N2, S2, SO2, CO2, SnS and SnO. These gases would usually be burned with stoichiometric air above the slag and pass through a waste heat boiler and fume collection system (as in conventional slag fuming) to an acid plant.

The matte gradually accumulates copper and precious metals and is tapped from time to time to bleed these metals from the system.

In an alternate mode of operation the ore is added continuously at a location as far as possible from the slag exit point. A sufficient volume of slag is maintained in the reactor to ensure a residence time such that fuming proceeds to the stage that the level of tin in the slag is sufficiently low for discard. The slag may be tapped either continuously or intermittently.

Although this process is autogenous when treating the ores for which it was developed, it is possible to supply extra heat by burning carbonaceous fuel in cases where the heat balance may not be favourable.

Irrespective of the mode of operation chosen, it is a simple matter to adjust the heat balance and the oxygen balance of the reactor by injection of mixtures of fuel and air in an appropriate ratio. Thus if the sulphur content of the matte becomes depleted but more gas needs to be swept through the system to ensure that fuming goes to completion, this gas and the required heat can be supplied by burning a carbonaceous fuel or pyrrhotite concentrate and adjusting the fuel/air ratio to balance the oxygen-to-sulphur ratio in the matte, thus maintaining the necessary conditions of stirring and gas stoichiometry for effective fuming.

Whether a sulphide or a carbonaceous fuel is used depends in turn on the iron-to-silica ratio in the feed to the process.

It will be obvious to those skilled in the art that the process described above is most closely related to the blast furnace pyritic smelting process described by Carlsson and Trostler.

However the present process differs in three important respects from the blast furnace process, namely: 1. The present process is a pneumatic matte conversion process conducted in a converter in which a bath of matte is continuously maintained and the air is blown through lances or tuyeres below the surface of the matte layer, which plays a predominant role in the chemistry, the heat transfer and the mass transfer processes taking place. In the blast furnace pyritic smelting operation it was not desirable to produce a matte, and the matte, if formed, did not play a prominent part in the reactions occurring in the bosh of the blast furnace.

2. In the blast furnace process it was necessary to carry out careful blending and preparation of the charge to promote conditions in the tuyere zone of the blast furnace such that combustion of the iron sulphides and formation of the slag could occur without causing bridging. In the present process the physical nature of the charge is not important. It is necessary only to crush the ore to a size suitable for easy handling and feeding onto the surface of the slag. Mixing of the charge with the matte and slag is assisted by the violent agitation in the bath.

3. In the earlier blast furnace process the matte, if it is produced, and slag, once formed, drip into the hearth below the tuyeres and are not subject to any further fuming reactions in the furnace. Both products must be collected and treated separately. The inventors claim a subsequent slag treatment stage and describe a separate treatment of matte if it should form.

In the present process the matte and slag are both subjected to the scrubbing action of a large volume of sulphidizing gas (generated by "conversion" of the matte and buffered chemically by the matte) during the entire residence time of the charge, and this reduces the tin content of both slag and matte to a low level and maintains a sulphidizing atmosphere around the charge particles. This buffering of the gas by its contact with a permanently maintained volume of matte distinguishes the present process from the prior art.

With regard to the rate at which the reactions occur, the present process may be likened to the new group of high-intensity processes which have been introduced for the smelting of copper, to replace the conventional blast furnace and reverberatory furnace, namely, the Noranda Process (see Extractive Metallurgy of Copper, Ed. Yannopoulis and Agarwal, AIM, New York 1975, Ch. 23, the Mitsubishi process (Ibid. chap. 22), both of which employ a type of converter. A recently published patent for continuous lead smelting also follows the same trend (United States Patent 3,663,207). Similar intense processes employing submerged combustion to supply heat and reductant have been described, e.g. zinc fuming (Australian Patent 429,266) and recovery of tin from slags (Australian Patent 465,531).

The following non-limitative examples illustrate features of the process: Example 1 300 grams of an iron sulphide matte of the approximate composition 67% iron, 29% sulphur, 4% oxygen was melted in an induction furnace in a non-oxidizing atmosphere. The furnace was instrumented to act as a reaction calorimeter that could measure net heats of reaction, with the intention of determining if the process described was autogenously heated.

When the temperature of the molten matte reached 12500C, 300 g of a pyritic tin ore, crushed to -4", was added slowly to the matte surface over a period of five minutes. Pyrite constitutes about 50% by weight of the ore composition. Other sulphides present are sphalerite, pyrrhotite, galena, chalcopyrite and stannite. Quartz and chert make up about 20% of the ore, other silicates present in minor amounts being chlorite, iron silicates and topaz. Carbonates which make up about 14% of the ore are represented by siderite, ankerite and huntite. Iron oxides make up about 4.0% of the ore, while small amounts of rutile, fluorite and apatite are also present. Cassiterite is present in an amount of about 2% by weight.

During the ore addition stage, some labile sulphur and stannous sulphide were evolved.

Once the ore had been added, air was admitted to the matte at a rate of 3 litres/minute via a ceramic lance that had been inserted in the bath. Vigorous oxidation occurred and the rate of tin sulphide evolution increased markedly. Minor amounts of sulphur were continuously evolved. The amount of air required was calculated on the basis that all iron in the ore would be oxidized to FeO, which would then combine with silica from the ore to form a fayalite slag.

In the present example, the ore analysed 26.3 % iron, from which 41.3 % iron sulphide could be formed.

In 300 g of ore, therefore, 300 x 0.413 = 123.9 gm of FeS = 1.40 moles /300 g ore.

The oxidation reaction is FeS + 14 O2 = FeO + 802 (5) therefore, 1.4 x 12 moles of oxygen is required to oxidize the available iron in the ore.

Amount of oxygen = 1.4 x 1.5 = 2.11 moles The amount of air is therefore: 2.11 x 0.21 = 10.05 moles At S.T.P., the volume required is 10.05 x 22.4 = 225.3 litres Therefore, at the addition rate of 3f/min., 75 minutes of oxidation would deliver the stoichiometric amount of oxygen required. This amount of air was then added to the bath and at the end of this time, the net heat of reaction was measured and found to be exothermic, therefore demonstrating the autogeneity of the process. The air was then stopped and the system cooled to room temperature under non-oxidizing conditions. The crucible was sectioned and the contents examined and analysed.Two distinct phases were present: 1) A matte phase that weighed 340 gm and analysed 26% S,2% SiO2, 65% Fe and 7% Oxygen, which was similar in composition to the original matte. The matte analysed 0.15% Sn.

2) A slag phase weighing 168 g which was silica saturated fayalite containing particles of silica and analysed 20% Si and 32%Fe. The tin analysis of the slag was 0.11%, thusenalbing an extraction efficiency to be calculated.

Results Weight % Sn Product gm % Wt %Sn Distribution Matte 340 56.7 0.15 5.3 Slag 168 28.0 0.11 1.9 Volatiles 92 15.3 * 92.8 Feed 600 100.0 1.6 100.0 * A sample of tin oxide analysed 59.2% Sn.

The above example illustrated that high extractions of tin can be obtained from pyritic tin ores using the sulphur in the ore as a heat source.

Example 2 300g of matte from the previous example, which contained 0.1% Sn, was used as the starting material for the second example. A similar procedure was followed and the same quantity of air added. The purpose of this experiment was to determine if any buildup or decrease of metals, such as copper, tin and gold occurred in the matte phase, and if any detrimental decrease of sulphur or iron occurred, that would change the exothermic character of the overall reactions.

Measurements indicated that the net heat of reaction was still exothermic, but slightly less than in the previous example.

No buildup of tin in slag or matte had occurred. Copper and gold concentrations in the matte had increased from 0.02 to 0.04% Cu and 3 ppm to 5 ppm Au respectively.

Results Weight % Sn Product gm % Wt %Sn Distribution Matte 302 50.3 0.15 4.7 Slag 207.4 34.6 0.11 2.3 Volatiles 90.6 15.1 - 94.0 Feed 600.0 100.0 1.60 100.0 Examples 3 & 4 Two more tests using recycled matte were conducted to show the effect of using matte that had been recycled four times.

The net overall heat of reaction was still exothermic, but only very slightly. In addition, the weight of matte had decreased, with a sympathetic increase in slag weight.

Concentrations of tin in the slag had decreased to 0.09% Sn in the case of example 3 and 0.07 % Sn in the case of example 4. Concentration of tin in the matte had decreased to 0.22% Sn after the fourth cycle.

Results (Test 4) Weight % Sn Product gm % Wt % Sn Distribution Matte 273.0 45.5 0.12 3.4 Slag 232.0 38.7 0.07 1.7 Volatiles 94.8 15.8 - 94.9 Feed 600.0 100.0 1.6 100.0 Copper and gold in the matte and built up to 0.1% and 9 ppm respectively.

The above experiments demonstrated that a batch-type system would be suitable provided that levels of tin in the slag and matte were acceptable and could be reduced if required. It was also clear that valuable metals could gradually be concentrated in the matte phase.

Example 5 Matte and slag from experiment 4 were both recycled and heated to 12500C under a non-oxidizing atmosphere.

The feed materials were therefore 250 g of matte assaying 0.12% Sn and 220 g of slag assaying 0.07 % Sn. In this experiment no fresh ore was added and 100 litres of air were added at 3e/min. to further oxidize the matte. The net heat of reaction was exothermic and the level of tin was reduced markedly in both slag and matte.

Results Weight % Sn Product gm % Wt % Sn Distribution Matte 194.1 41.3 0.04 16.5 Slag 215.3 45.8 0.03 13.7 Volatiles 60.6 12.9 - 69.8 Feed 470.0 100.0 0.10 100.0 This test showed that any buildup in tin in slag or matte could be overcome by addition of extra air. However, the gradual loss in weight of matte may eventually result in the reactions being net consumers of heat, with the periodic make-up of fresh matte then being an essential requirement in batch type processes.

Example 6 Using 300 g of a fresh matte sample, and 300 g of ore, the air flowrate was doubled to 6 litres/min. Consequently, oxidation time was halved to 37 minutes.

Results: were as follows Weight % Sn Product gm % Wt % Sn Distribution Matte 354.0 59.0 0.12 4.4 Slag 166.2 27.7 0.07 1.2 Volatiles 79.8 13.3 - 94.4 Feed 600.0 100.0 1.60 100.0 This test showed that equally good extraction could be obtained using double the air addition rate. The net reaction was measured as exothermic.

'Micrographic examination of the slags and mattes from the above series of tests showed that a considerable quantity of magnetite was present in all samples.

Example 7 Experiment 1 was repeated using a finer sized ore feed (-12 + 30# B.S.S.) to indicate if reactions were increased with a finer crushed ore. Results show that an improvement in recovery is possible with finer feed.

Results Weight % Sn Product gm % Wt % Sn Distribution Matte 353.4 58.9 0.08 3.0 Slag 156.6 26.1 0.03 0.5 Volatiles 90.0 15.0 - 96.5 Feed 600.0 100.0 1.60 100.0 Example 8 As the process had been successfully tested using a pyritic tin ore, it was decided to test a different type of ore. The ore selected was a pyrrhotitic tin ore of the following approximate composition: 1.2% Sn 36.0% Fe 22.0% S 0.9% C02 21.0% SiO2 The mineralogical composition of the ore was: 55 - 60% FeS (pyrrhotite) 1 - 2% SnO2 (cassiterite) 3% FeCO3 (siderite) 21 - 25% sio2 (quartz) 5% Iron Oxides 5% Gangue and other minerals such as fluorite, dolomite and iron silicates.

The cassiterite was mainly combined with quartz and was considerably coarser in grainsize than the pyritic tin ore. The ore was crushed and screened to -121F + 36# B.S.S. The stoichiometric amount of air required was calculated on the basis of 22.0% S in the ore.

Therefore, in 300 g based on sulphur analysis, 60.5% FeS is possible = 2.06 moles.

Oxygen required = 3.09 moles Air required = 3.09 22.4 0.21 x 1 = 330 litres at S. T.P.

A similar experimental procedure to that described in example 1 was used. Once again labile sulphur and stannous sulphide were evolved prior to and during addition of air. SO2 was also detected in the off gases. The net reaction was very exothermic. This was principally due to the extra heat evolved from the extra available pyrrhotite, and elimination of heat required to remove the labile sulphur from pyrite.

Extraction of tin into the volatile fumes was comparable to the pyritic ore.

Results Weight % Sn Product gm % wt % Sn Distribution Matte 427.2 71.2 0.10 5.9 Slag 139.8 23.3 0.08 1.5 Volatiles 33.0 5.5 - 92.6 Feed 600.0 100.0 1.20 100.0 This test demonstrated that pyrite was not essential for the success of the process and that satisfactory tin extraction could be achieved from a pyrrhotitic tin ore. However, the increase of matte in the process indicated that more oxidation air, or alternatively more silica, would be required to balance the Fe/Si ratio in the slag.

Example 9.

A third ore type, largely siliceous, was tested. The ore analysis was: 0.98% Sn 4.0% S (as pyrite and pyrrhotite) 15.0% Fe 29.0% Si 2.0% CO2 The ore mineralogical composition was approximately 1.3% SnO2 7-8% FeS2 and FeS 60% SiO2 4% FeCO3 10-15% Iron oxides 5% Fluorite 5-19% Other minerals 100.0 % Ore size was 12# B.S.S.

The test procedure was similar to Example 1. 300 g of matte was melted and maintained at 1250"C, to which 300 g of the siliceous tin ore was added and the stoichiometric amount of air added to oxidize all sulphur in the ore, assuming that all of it sulphidized some iron to iron sulphide. This was calculated at 41 litres of air and this was bubbled through the bath at 36/mien. Once again, stannous sulphide and sulphur vapour were evolved, but after 5 minutes of oxidation, the amount of sulphur decreased in favour of an increase in sulphur dioxide.

The net heat of reaction was endothermic due to the low sulphur levels in the ore. However tin extraction was reasonably good.

Results Weight % Sn Product gm % Wt % Sn Distribution Matte 308.4 51.4 0.23 12.0 Slag 229.1 38.1 0.12 4.7 Volatiles 62.9 10.5 - 83.3 Feed 600.0 100.0 0.98 100.0 This indicated that a largely siliceous tin ore, of relatively large cassiterite grainsize could be treated using the matte fuming technique. However, a fuel source such as pyrrhotite or carbonaceous fuel would be required to provide heat. The pyrrhotitic tin ore described above would be a suitable fuel and fluxing agent for the siliceous ore. Alternatively, the siliceous ore could be added to the pyritic or pyrrhotitic ores to aid fluxing and adjust the iron-silica ratio of the feed to suit process requirements.

Example 10 The next logical step was to blend the pyrrhotitic and siliceous ores to present a suitable feed from both a heat balance and fluxing viewpoint.

A mixture of 75% pyrrhotitic ore and 25% siliceous ore resulted in a feed assaying: 1.15% Sn 17.5% S 30.8% -Fe 16.5% Si The size of the composite ore was - 12# B.S.S. Using the normal procedure, it was calculated that the stoichiometric amount of air required (based on 17.5% S) was 262.5 litres. This was bubbled through a mixture of 300 g of ore and 300 g of matte and the net heat of reaction was measured as exothermic. The fluidity of the slag was noticeably better than any previous experiment, and fuming rates of tin sulphide and sulphur were substantial.

Results weight % Sn Product gm % Wt % Sn Distribution Matte 373.2 62.2 0.09 5.0 Slag 171.6 28.6 0.15 3.8 Volatiles 55.2 9.2 - 91.2 600.0 100.0 1.15 100.0 This experiment demonstrated that by blending ores to obtain a suitable feed from both a slag-forming and heating viewpoint, good extractions of tin were possible and reactions could be made exothermic.

Example 11 Previous tests demonstrated the features of the process on a batch basis, by recycling cold matte.

However, the gradual decrease in matte weight indicated slight overoxidation due to the distillation of sulphur when adding ore. Consequently, less than the theoretical amount of sulphur was available as fuel, so it was decided to reduce the amount of air by 10%, the aim being to maintain a constant matte weight and autogeneous heating.

Two kilograms of iron sulphide was melted under nitrogen in a crucible and the tempera ture maintained at 12300C. Air was added at 10 litres/min. to the bath and 2 Kg of pyritic tin ore added as 200 g batches every two minutes. The calculated amount of air, 1600 1, was added and the matte and slag then allowed to settle for ten minutes under nitrogen cover gas.

Slag was then poured from the crucible via an attached spout and the matte and a thin slag layer retained in the crucible as the starting matte for the next cycle. This initial slag analysed 0.12% tin.

In all, the above cycle was repeated eight times and after the final cycle, matte was also poured from the crucible. Clearly, this test was a more realistic demonstration of the process capabilities under semi-continuous conditions. All slags and matte were weighed and assayed for tin, sulphur and iron.

The amount of slag tapped varied from 30% to 40% of the weight and the tin level decreased from 0.12% Sn to 0.0s Ojo Sn after three cycles and levelled off at about 0.06-0.08% Sn for the remaining cycles. Losses to slag were always less than 5% of the tin distribution.

Sulphur levels in slag were always less than 4% S and iron levels between 30-40%.

The matte assayed 0.09% Sn, 64.1% Fe and 20.9% S and represented 52.3%of the weight, thus showing that the matte bath could be maintained over a long run. Less than 4% of the tin from the last cycle was retained in the matte.

Overall, the heat balance was slightly exothermic indicating that the process was potentially autogeneous.

Example 12 A large scale test was carried out in which 90 Kg of pyritic tin ore (0.93 % Sn) was treated in a 50 Kg capacity refractory lined vessel similar to that described by J.M. Floyd in Paper 3.5, Fourth World Conference on Tin, Kuala Lumpar 1974. Oil to cover furnace heat losses and air sufficient for oil and ore oxidation were delivered into the bath via a lance, thus utilizing the submerged combustion technique to overcome heat losses. Ore crushed to -a" was fed onto the surface of the bath of 30 Kg/hr for the first hour and then at twice the rate for the remainder of the test.Significant heat was generated and the temperature was maintained above 1300"C, and approached 1400"C when the addition rate was doubled, indicating that the process was close to being autogeneously heated.

Two quantities of slag were tapped and the fume was collected after the test. Over 95 % of the tin was extracted and slag tin levels were lower than in any previous tests, indicating that continuous, small increment feeding was at least as good as batch feeding. The rate of tin elimination was obviously very fast.

Maximum tin concentration in slags was 0.09% Sn and the average slag assayed 0.063%sun.

Fume analysed 45% Sn and was contaminated to some extent by ore and slag dust.

Example 13 A second large scale test was conducted in which the ore, crushed to -3/8" was screened through a 10 mesh BSS screen. The fine fraction amounting to 20% of the feed was injected into the matte bath through the lance, while the coarse fraction was fed at a constant rate onto the surface of the molten bath. In other respects the test was identical with example 12.

The maximum tin concentration in the slag was .08% Sn and average slag assayed 0.05% Sn.

The fume from this test was noticeably cleaner, and contained 55% Sn.

This example shows the improvement resulting from the avoidance of loss of fines during feeding of the ore.

In the following claims the term "ore" is to be understood as including concentrates or other mineral mixes.

WHAT WE CLAIM IS: 1. A process for extracting tin values from a tin bearing iron sulphide ore, comprising charging said ore in particulate form into a refractory-lined reactor containing a pool of molten matte so that said iron sulphides in said ore become molten and enter the matte pool, blowing air or other oxidizing gas through the matte pool to oxidize said iron sulphides thus producing iron oxide and generating the heat required to smelt the charge and drive off the tin values from the charge in the form of volatile sulphides and oxides, said iron oxide combining with silica added in the charge to form a layer of molten slag which floats on the matte bath, the conditions in the reacting mass being such that the oxidizing gas blown into the reactor comes first into contact with the pool of molten matte, thus stabilizing the oxygen and sulphur pressures of the gases, after contacting the matte, in a suitable range to sulphidize and volatilize the tin in the charge without the need to add a carbonaceous reductant.

2. A process according to claim 1, wherein additions of iron-bearing or siliceous materials are made in order to adjust the iron-to-silica ratio in the slag.

3. A process as claimed in claim 2, wherein pyrrhotite is added to adjust the iron-to-silica ratio in the slag.

4. A process according to any preceding claim, wherein, after the completion of the ore addition, blowing is continued until the stoichiometric oxygen has been supplied to convert all the new iron sulphide matte resulting from the addition of the charge of ore.

5. A process according to any preceding claim, wherein fuel is burnt to offset heat losses from the reaction.

6. A process according to claim 5, wherein the fuel is a carbonaceous fuel.

7. A process according to claim 5, wherein the fuel is pyrrhotite or pyrite.

8. A process according to any preceding claim, wherein the gases containing the volatile sulphides are burnt and the tin values collected in a fume collection system.

**WARNING** end of DESC field may overlap start of CLMS **.

Claims (12)

**WARNING** start of CLMS field may overlap end of DESC **. poured from the crucible. Clearly, this test was a more realistic demonstration of the process capabilities under semi-continuous conditions. All slags and matte were weighed and assayed for tin, sulphur and iron. The amount of slag tapped varied from 30% to 40% of the weight and the tin level decreased from 0.12% Sn to 0.0s Ojo Sn after three cycles and levelled off at about 0.06-0.08% Sn for the remaining cycles. Losses to slag were always less than 5% of the tin distribution. Sulphur levels in slag were always less than 4% S and iron levels between 30-40%. The matte assayed 0.09% Sn, 64.1% Fe and 20.9% S and represented 52.3%of the weight, thus showing that the matte bath could be maintained over a long run. Less than 4% of the tin from the last cycle was retained in the matte. Overall, the heat balance was slightly exothermic indicating that the process was potentially autogeneous. Example 12 A large scale test was carried out in which 90 Kg of pyritic tin ore (0.93 % Sn) was treated in a 50 Kg capacity refractory lined vessel similar to that described by J.M. Floyd in Paper 3.5, Fourth World Conference on Tin, Kuala Lumpar 1974. Oil to cover furnace heat losses and air sufficient for oil and ore oxidation were delivered into the bath via a lance, thus utilizing the submerged combustion technique to overcome heat losses. Ore crushed to -a" was fed onto the surface of the bath of 30 Kg/hr for the first hour and then at twice the rate for the remainder of the test.Significant heat was generated and the temperature was maintained above 1300"C, and approached 1400"C when the addition rate was doubled, indicating that the process was close to being autogeneously heated. Two quantities of slag were tapped and the fume was collected after the test. Over 95 % of the tin was extracted and slag tin levels were lower than in any previous tests, indicating that continuous, small increment feeding was at least as good as batch feeding. The rate of tin elimination was obviously very fast. Maximum tin concentration in slags was 0.09% Sn and the average slag assayed 0.063%sun. Fume analysed 45% Sn and was contaminated to some extent by ore and slag dust. Example 13 A second large scale test was conducted in which the ore, crushed to -3/8" was screened through a 10 mesh BSS screen. The fine fraction amounting to 20% of the feed was injected into the matte bath through the lance, while the coarse fraction was fed at a constant rate onto the surface of the molten bath. In other respects the test was identical with example 12. The maximum tin concentration in the slag was .08% Sn and average slag assayed 0.05% Sn. The fume from this test was noticeably cleaner, and contained 55% Sn. This example shows the improvement resulting from the avoidance of loss of fines during feeding of the ore. In the following claims the term "ore" is to be understood as including concentrates or other mineral mixes. WHAT WE CLAIM IS:

1. A process for extracting tin values from a tin bearing iron sulphide ore, comprising charging said ore in particulate form into a refractory-lined reactor containing a pool of molten matte so that said iron sulphides in said ore become molten and enter the matte pool, blowing air or other oxidizing gas through the matte pool to oxidize said iron sulphides thus producing iron oxide and generating the heat required to smelt the charge and drive off the tin values from the charge in the form of volatile sulphides and oxides, said iron oxide combining with silica added in the charge to form a layer of molten slag which floats on the matte bath, the conditions in the reacting mass being such that the oxidizing gas blown into the reactor comes first into contact with the pool of molten matte, thus stabilizing the oxygen and sulphur pressures of the gases, after contacting the matte, in a suitable range to sulphidize and volatilize the tin in the charge without the need to add a carbonaceous reductant.

2. A process according to claim 1, wherein additions of iron-bearing or siliceous materials are made in order to adjust the iron-to-silica ratio in the slag.

3. A process as claimed in claim 2, wherein pyrrhotite is added to adjust the iron-to-silica ratio in the slag.

4. A process according to any preceding claim, wherein, after the completion of the ore addition, blowing is continued until the stoichiometric oxygen has been supplied to convert all the new iron sulphide matte resulting from the addition of the charge of ore.

5. A process according to any preceding claim, wherein fuel is burnt to offset heat losses from the reaction.

6. A process according to claim 5, wherein the fuel is a carbonaceous fuel.

7. A process according to claim 5, wherein the fuel is pyrrhotite or pyrite.

8. A process according to any preceding claim, wherein the gases containing the volatile sulphides are burnt and the tin values collected in a fume collection system.

9. A process according to any preceding claim, wherein the matte is tapped from time to

time to recover copper and/or other metals which accumulate therein.

10. A process according to any preceding claim, wherein the ore is added continuously and a sufficient volume of liquid iron silicate slag is maintained in the reactor to ensure a residence time such that fuming proceeds to the stage that the level of any tin that may enter the slag is sufficiently low for discard.

11. A process according to any preceding claim, wherein the air is blown through tuyeres or lances below the surface of the matte layer.

12. A process for extracting tin values from a tin-bearing ore substantially as hereinbefore described in any one of the Examples.