Ferrosilicon powder (containing approx. 15% Si) for constitution of heavy media is used today in many separation plants treating iron and other ores. The production of the powder by grinding is expensive. A new process atomizes the still molten ferrosilicon by air or steam.

In a heavy medium this atomized ferrosilicon powder has several remarkable properties: The round form and smooth and shining surface of the individual grain offers greater resistance to corrosion, especially as no contamination is produced between iron and silicon during the process. The viscosity of the heavy media pulp for the same weight of pulp is less than for ground ferrosilicon, i. e. for the same viscosity a greater pulp density can be attained. This might be of importance in future. Adhesionloss is lower for atomized ferrosilicon. The strength of the atomized powder is equal to that of the ground powder.

Ferrosilicon as a Heavy Medium

The use of ferrosilicon powder has risen with the increasing employment of heavy media separation for ores, particularly in the separation of iron ores. In the separation of lead and zinc ores, ferrosilicon is a foreign body and the use of galena as a heavy medium would seem indicated; yet ferrosilicon has largely replaced galena for this work. The reason for it is that ferrosilicon possesses good magnetic properties which can be used to recover it from fouled medium. Galena can only be recovered by flotation. In contradistinction to other heavy media like magnetite, ferrosilicon has the advantage of a high specific gravity and of greater hardness. In America about 15-20 million tons of iron ores were separated with ferrosilicon in 1952.

A powder with about 15% Si=content is used because in it the magnetic properties are well distinguished and the specific gravity is high. With a Si=content of 18-20% the spec, gr. is lower and the limit of magnetizability is near. With a content of only 12-14% Si resistance to corrosion is lessened. The 15 % ferrosilicon is melted in a special electrical furnace. The desired Si=content is obtained by melting down a suitable mixture of scrap, gravel and carbon. It is, of course, possible to start with the available low percentage ferrosilicon and to add gravel and carbon, or to blend high percentage ferrosilicon, such as is obtained from the carbide furnace operation, with scrap only. If a considerable amount of slag occurs during the process, the standard electrode- furnace is more suitable than an induction furnace, because the lining of the latter is easily corroded by slag. The melt in the furnace reaches a temperature of approx. 1300° C (2350° F).

Ferrosilicon Powder Manufacturing Process

By Grinding

In the standard process liquid ferrosilicon is removed from the furnace, cooled, crushed, and ground to the desired fineness. Fine and coarse powders are available. The finer one is preferred in the first stages of a sink and float plant. After some running time the coarser quality is added to make up the loss and to correct wear. The actual separating vessel is of great importance; it should be so formed that the heavy sink product at its bottom cannot grind the ferrosilicon as in a ball mill.

In many cases wet grinding is preferred as the less expensive process. This, however, is done locally because the ground sludges cannot be transported easily. Any drying must be carried out with extreme care because at higher temperatures decomposition occurs in consequence of increased corrosion. These grinding plants are rather expensive, and are often not running to their full capacity; yet it is useless to build a smaller plant for the mills must have a certain size for an economical treatment of this material. Ferrosilicon with a higher Si-content (for example, 45 or 75% Si) can be ground more easily because it is more brittle than the 15% which, as has been shown in tests, possesses a certain toughness making comminution difficult. Tests have shown that the grinding operation causes great wear of the mills and is costly.

By Atomizing

A new method of powder production has therefore been developed whereby the liquid melt is atomized by air or steam.

The resulting particles are quenched in water, and the powder which after draining still contains approx. 10 % water is dried and screened. In contradistinction to the ground powder, the atomized powder shows neither decomposition nor rusting. This process is similar to the production of metal powders used in powder metallurgy. Provision must be made for substantial supplies of steam at controlled pressure. Approx. 2 tons of steam are used per ton of ferrosilicon.

Atomization also produces finer or coarser grains as desired. The determinants are the steam pressure, temperature and composition of the ferrosilicon melt. It has been found to be of special advantage if, besides copper, other alloys are present in the melt, in the concentration needed to produce a good powder by atomization. As with the ground ferrosilicon so with the atomized ferrosilicon the size distribution in the Rosin-Rammler diagram is represented by a straight line (figure 1), which was not quite expected.

Properties of Atomized Ferrosilicon

The two different processes of operation (grinding and atomizing) result in different properties in these products, of which only those interest us which influence the heavy media pulp. Decisive for the different behaviour is not so much the chemical composition which is the same in both cases, but the physical condition of the surface of each individual powder particle (figure 2). The ground powder shows sharp cor-

ners and edges and has a rough surface, whereas in the atomized material the individual grain has a more or less rounded form as well as a smooth, shining surface which is specially hard, due to quenching. As mentioned above, the formation of this surface depends chiefly on the composition of the melt.

This difference in the surface determines the different behaviour of the two powders with respect to

1. corrosion,
2. viscosity of pulp, and
3. loss through “drag=out” or adhesion, ferrosilicon to material treated.
   These three items determine the cost of this medium in a sink and float plant. There is further
4. the properties of strength.

1. Corrosion

It is known that under the influence of water the ground powder, after some time, begins to give off hydrogen and to rust. When wetted repeatedly with water and allowed to dry between wettings, microscopical examination shows that the powder begins to rust. A quick test procedure for ferrosilicon powder has been worked out; the powder, together with plenty of water, is put into a beaker, and treated in a waterbath at 70° C.

According to the susceptibility to corrosion the beginning of gas emission can be observed at different time intervals. This quick test procedure has also shown that even 45% and 75% ferrosilicon powder reacts with water, if it is fine enough for the requirements of sink and float separation. Accurate tests have shown that the development of hydrogen generally starts at the corners and edges where according to the latest theories in surface chemistry rust-forming influences are present which are very active. Therefore corrosion cannot be avoided by chemical additions, not even copper, because it depends on the surface conditions. Another point is that during the grinding process iron particles are worn off the grinding bodies and the mill lining. Such iron particles which are either mixed with or rolled onto ferrosilicon particles form with the latter so-called “local elements” or electro couples which strongly promote the formation of rust. Such local elements are also formed within the individual powder particles if several chemical compounds between the iron and silicon segregate as the tapped melt cools down slowly. All these compounds have different potentials, and therefore give rise to the formation of smallest elements. Measurements of these potentials showed that for example there is a potential difference of 300 Millivolt between 15% ferrosilicon and iron in a 3% sodium chloride solution. In practical operation the corrosion results in agglomeration of the settled powder along with a more or less heavy formation of gas. Hydrogen is formed as can be seen from the following equations:

Si + 2 H2O = SiO2 + 2 H2

2Fe + H2O + O2 = Fe2O3 + H2

SiO2 and Fe2O3 occur as hydrates and cause cementing of the deposited particles. Though in the decomposition of iron the hydrate of the three valency iron oxide (Fe2O3 · xH2O) is not formed primarily, the primarily formed hydrate of the two valency iron FeO · xH2O changes in the presence of oxygen quickly to the three valency form. Oxygen is always present as the pulp is well aerated by the strong movement. Many heavy media plants use air for the lifting of the liquid. The atomized ferrosilicon behaves differently. Here, the individual particle does not offer preferential points of attack to rusting as its shape is spherical, and its surface hard. Due to the quenching, segregation of various chemical compounds does not occur and contamination by worn-off iron does not exist. In the quick test procedure the formation of gas bubbles is strongly reduced. It occurs only after 6-7 hours whilst in ground material it can be observed within ½ hour. Therefore in operation agglomeration of the settled powder does not occur. Experience teaches that even after shut-downs over the weekend sink-float plants can be restarted without any trouble. After a run of 20 to 30 minutes everything is in suspension, and the feeding of ore may be started. Of course, the formation of gas can be avoided only with pulps which are not acid. Acidity may result when sulphidic ores are being treated. In this case hydrogen-sulphide is formed which lowers the pH. Continuous control of the latter is then necessary to prevent it from dropping below 6.3. In such cases addition of small amounts of lime or washing soda is necessary. Loss through corrosion, for ground ferrosilicon up to 10-15% of the total loss, decreases considerably when atomized ferrosilicon is used.

2. Viscosity of Pulp

Only as long as the viscosity does not exceed the tolerable limit, is the heavy liquid serviceable and suitable for separating relatively small-sized feed. When plotting the viscosity against the pulp density the range of serviceability is confined to the horizontal part and the first third of the rising part of the curve. Lacking a suitable measuring procedure for suspensions (non-Newtonian liquids) the viscosity was determined with a run-out viscositymeter. The graph (figure 3)

shows the influence of particle shape on viscosity. It has long been known that for the same pulp density and the same screen size a spherical shape causes lower viscosity than material with corners and edges. The forces which cause particles to adhere to each other depend on the effective adsorption and friction between the surfaces of such particles. It is obvious that with particles of rough shape where the surface is made up, more or less, of actual planes which, however, are at the same time rough in structure, adhesion will occur to a much greater extent than is possible with ball-shaped particles the surfaces of which are smooth and reflecting, and which offer conditions of ball-contact only. With atomized ferrosilicon the particle surface is still smoother than with originally cornered material which has been worn, in the course of time, to nearly spherical shape. The latter has also been referred to as having spherical shape, but the particle surface is not shining and hard as with a powder produced by atomization. For the same viscosity of approx. 12 centipoise, with cornered material according to the graph a pulp density of 3.3 can be attained, with atomized powder 3.45, and with a special grade consisting of spheres only, a pulp density of 3.7 can be reached, as has been shown in the second picture. In practice a density from 3.1 to 3.2 only is attainable with normally ground ferrosilicon whereas 3.45 is reached with atomized material, and 3.9 with special grade ferrosilicon. As regards the last mentioned high pulp density operating results from larger plants are as yet not available; a suspension of 3.45 density, however, is being used in an extensive plant. Thus, the pulp densities so far attained in practical operation can be substantially increased without increasing the viscosity. Vice versa, with equal values of pulp density the use of atomized ferrosilicon will result in lower viscosity values than can be reached with the ground material. On this basis it may in future be possible to solve certain problems in mineral dressing which so far have been intractable, as suitable pulp densities were not attainable. For individual cases increase of pulp density will result in a sharper separation, and it is conceivable that, for example, fluorspar and barytes can be separated at a pulp density of 3.8.

3. Adhesion — Losses

Due to the spherical shape and smooth surface of atomized ferrosilicon particles, the heavy media losses in practical operation are reduced. Such losses are due to the fact that not all of the ferrosilicon powder that adheres to the ore or gangue after treatment is removed by spraying the sink and float products on the rinsing screens. It is obvious that a more or less ball shaped grain of powder rolls off a product particle more easily than does a cornered grain. In this respect the surface structure of the ore or gangue is another important factor. With smooth product surfaces the losses in ferrosilicon are, of course, lower than with rough ones. By model tests it was possible to show that, as to adherence losses, savings from 20 to 60 % are obtained. These tests were carried out as follows: 100 g ore of a size fraction of 20—10 mm were put into the pulp for a short period, allowed to drain on a vibrating screen and for five seconds moved under a spray. The individual pieces were taken from this to another screen, and the rest of ferrosilicon washed off with clean water. Having repeated this 10 times, the water was filtered off, and the ferrosilicon weighed. All conditions proved to be reproducible. The comparing measurements were carefully carried out by the same technician. Individual figures for adherence losses, in grammes per ton, are shown in table 1.

4. Strength Properties

It might be considered a disadvantage of the atomized powder that especially in the coarser fractions the spheres or spherical shaped grains contain cavities which can be noticed under the microscope after particle destruction, and into which the pulp can penetrate. However, this has no bearing on its property as a heavy medium. If this property affected the work it would be noticeable immediately in a decrease of the true specific gravity, and it would be impossible to attain the above mentioned high pulp densities. There is only a small difference between the original powder and such as has been further comminuted through grinding resulting in destruction of the balls; from this the ratio of closed pores has been found to be only 1.8%. The reason for this is that only the coarser fractions of the powder, and of which there exists but a small amount, show the mentioned cavities whereas the main powder fraction consists of < 0.1 mm grains. The latter contain no cavities, but are solid throughout. It is conceivable that the cavities might have a further detrimental influence upon the strength of the individual particles. On the other hand it is known that a hollow shape has a greater compressive strength than a solid body (a pipe has a greater strength than a rod of the same diameter). Abrasion tests were therefore carried out with ground and atomized ferrosilicon powder in a small ball mill. Equal size fractions of 2-0 mm were ground for several hours under uniform conditions, and the —60 micron fraction determined after

different grinding periods. The graph (figure 4) shows that after a short period (½ hour) the ground ferrosilicon showed increased abrasion, whereas after a longer period the atomized ferrosilicon yielded a greater amount of fine material. A chosen fraction of the total material, for example 0.09— 0.06 mm (figure 5) shows greater abrasion for the ground than for the atomized powder; however, the differences of these measurements are small and may well be within the tolerance of experimental errors. From these abrasion tests it may be concluded that the strength of the atomized powder is not less than that of the ground powder.

To sum up, it may be stated that the good results experienced up till now with ferrosilicon pulps of ground powder have been improved with the atomized powder. This is proved by the operational results of various separation plants over a period of several years. The relevant journals have from time to time published further information.