Zinc Flotation Concentrate Roasting Furnace

This paper describes experiments carried on at the Case School of Applied Science, together with their results. Their success led to the design of the larger furnace herein described, but which has not been built.

A previous article by the authors contained a general description of the new roasting furnace herein described but it did not go into detail as to the metallurgical behavior or the results obtained. Believing that such information would be of great value, they have elaborated on the subject and have given many unpublished details.

The furnace described applies the principle of roasting finely divided zinc-sulfide ores, now produced in large quantity by the flotation process, in gaseous suspension; that is, the ore particles are carried, in suspension, in a current of air and gaseous products of the roast. The relatively great fineness of flotation concentrate presents difficulties and problems of roasting in furnaces of the ordinary type; the fine ore is forwarded through the furnace in the form of a shallow bed and its very fineness leads to dense impervious bedding which prevents oxygen from reaching the interior of the bed, thus unduly lengthening the time of roasting and preventing the elimination of the last of the sulfur. The fineness of the concentrate, normally, should lead to a rapid and complete roast, for the speed of roasting is a function of the surface exposed to oxygen, which surface is greatest, per unit of weight, in very fine material. The difficulty in bed roasting is to get the oxygen to the particle. If the fine ore, during the roasting, could be freely suspended in oxidizing gases, full advantage could be taken of the great surface conferred by its fineness. This fundamental idea, of course, is not new, for the Stedtefelt furnace, familiar to the older metallurgists, is an example of it; but the manner in which this is accomplished may be new.

The numerous efforts to roast in gaseous suspension show that the idea is attractive; in fact, an analysis would indicate that it is the most reasonable way to effect the oxidation of ore, provided certain difficulties can be overcome. Two objections against such a method that formerly had much weight were the cost of fine grinding the ore, also the fact that the finely ground product, even after roasting, was not the best condition of material for further metallurgical operations. The objection of costly grinding, however, has been removed by the production of great quantities of flotation concentrate, which in point of fineness present a material that is ideal for roasting in gaseous suspension.

The inception of the experiments described here is due to David B. Jones and March F. Chase. The idea of the general type of furnace and process was suggested to the authors and the experimental work was carried out in the metallurgical laboratories of the Case School of Applied Science at Cleveland in 1915-16. The original plan was to make a furnace for roasting zinc-blende flotation concentrate that would use the heat value of the sulfide to accomplish the roast and to produce a gas suitable for making sulfuric acid; i. e., of sufficient concentration in SO2 and practically free from the products of carbonaceous combustion. Autoroasting of sphalerite is theoretically possible, and also practically, as was demonstrated in the experiments set forth.

A diagrammatic drawing of the roasting furnace and accessory apparatus as erected is shown in Fig. 1. Here A is the furnace proper; B the two stoves, heated by natural gas, that preheat the air; and C, the Cottrell electric precipitator for the precipitation of flue dust and fume carried from the furnace by the gases, which in a commercial plant would pass to a sulfuric-acid plant. At H is the gas supply for the stoves; X is the dust chamber, and G is the cycloidal blower that furnishes air to carry the ore in suspension in the furnace. The dried, preheated ore (60° to 100° C.) is charged into the hopper 1 whence the endless screw 2, operated by a variable-speed electric motor 3, discharges it into a pipe 4, directly above the nozzle 5. A stream of high-pressure (20 to 60 lb.) moderately pre-heated air is discharged through this nozzle in such quantity as to carry readily the fine ore in suspension into pipe 7, on the injector principle. Pipe 7 is of larger diameter than pipe 4 and is lined with refractory material; it serves as the main injector pipe into the furnace. Air, heated to approximately 800° C., passes from one of the stoves B to the injector pipe 7, through the supply pipe 9; this is the main air supply for roasting the ore.

The amount of air supplied is governed by two principles: (1) The quantity must be correct to roast the ore and to furnish a gas of the correct composition for the manufacture of sulfuric acid. Every pound of

sphalerite requires 35.7 cu. ft. of air (standard conditions) to convert the zinc to oxide, the sulfur to dioxide, with enough more oxygen to convert this into the trioxide. Some excess must be carried; in the experiments from 41 to 55 cu. ft. and sometimes 75 to 100 cu. ft. were used, as measured by a meter. (2) There must be such a relation between the quantity of air per minute and the area of the riser tube 10, that the velocity of the ascending air current will be enough to carry the largest ore particle to the top of this tube and over the edge. Fig. 2 gives definite data on this point for sphalerite. Fortunately the requirements for both conditions can readily be fulfilled.

The mixture of ore and high-pressure air from pipe 4 enters the main injector pipe 7 with a rotary motion and is caught by the ascending hot

air from the pipe 9 and injected into the riser tube 10. The fundamental idea is to have the temperature of the ore-air mixture at the ignition point of zinc sulfide, which varies between 650° and 810° C., depending on the size of the particle as it leaves the injector tube to enter the riser tube, so that the full time of the ore particle in the furnace will be available for oxidation. After leaving the riser tube 10, the ore particle falls in the annular space 11, the area of which is large so that the natural velocity of fall shall not be augmented by the velocity of the descending gases. It is the belief of the authors that the gas envelope surrounding the ore particle is constantly changing; thereby causing fresh oxygen to be supplied to the particle.

The time necessary for the roast is furnished by the passage of the ore up the combustion tube 10 and its fall in the annular space 11. This time depends on the height of the furnace, the length of the path of travel being practically twice’ the height of the furnace, also on the velocity of the ascending air current in the combustion tube. Assuming a definite ratio of cubic feet of air per pound of ore, the velocity will be determined by the area of the combustion tube. Definite figures on this point are given later. In any given furnace, i. e. a fixed area of combustion tube, there is a certain variation allowable in the velocity of the ascending gas current, obtained by varying the amount of air, which will then vary the time of the ore in the combustion tube. Too high a velocity, obtained either by too much air only or too great an ore feed with its corresponding increased amount of air, will shorten the time element so much that the sulfur is not sufficiently eliminated. The greater part of the ore collects in the hopper 14; the gases and fine dust pass, through openings 12, into the flue 13, thence to the settling chamber X, where the coarser dust is settled out, thence to the Cottrell precipitator C for the precipitation of the finest dust and fume.

Thirty-Foot Experimental Furnace

The design of the 30-ft. experimental furnace was, as far as possible, based on the results obtained in the previous work but was limited by lack of space and the capacity of the stoves, blowers, gas supply, etc. The stoves had been erected for the small furnace already described and were known to be inadequate for the larger furnace, but no space was available for enlarging them. It was also desirable to keep the expense as low as possible, so that some things that would have aided in the work were omitted.

The furnace consisted essentially of a brick stack (Fig. 3) approximately 30 ft. in height from the base plate to the top of the cover arch. The base plate and support rested on 4-ft. reinforced concrete posts, thus making the total height about 35 ft. The space between the posts provided room for the calcine hoppers and air and feed inlets and a pit provided the additional space necessary for the cleaning of tools and for making repairs.

The stack was built of two concentric firebrick circles 24 and 36 in. internal diameter; the space between these circles was filled with mineral wool. Somewhat larger circle brick were used near the bottom of the stack to increase the stability. No steel shell was placed outside of the brick, though such a shell would have aided greatly in the construction

of the furnace and in carrying on the tests. The top of the stack was closed by fireclay slabs, a small circular hollow tile being set at the center; a gate at its top permitted the interior of the furnace to be watched while operating. Two hoppers at the bottom of the stack collected the greater part of the roasted ore. Above these hoppers, and spaced at equal intervals around the stack, were four openings to the flue that led to the settler and to a small Cottrell precipitator.

Concentrically with the walls of the stack was placed the riser, or combustion tube, which extended from below the hoppers to within about 5 ft. of the furnace top. This tube, in the earlier trials, was made of

9-in. hexagon stove tile but later 12-in. cylindrical tile was used. The annular space between the combustion tube and the stack walls served as a downtake for the products of the roast.

The injector nozzle entered the bottom of the combustion tube for about 5 ft., thus placing its delivery point about 2½ ft. above the base plate of the furnace. The nozzle was a 2-in. fireclay tube and connected at its lower end with a pipe or mixer, into which the air and ore were injected from the stoves and feeder respectively.


The two stoves were of the central combustion type, the combustion tube being made of 9-in. hexagonal stove tile. The checkers were made of 2 by 2 by 9-in. brick laid to provide the maximum heating surface. A 1/8-in. steel shell with top and base plates insured airtight conditions. The stoves were 4 ft. in diameter, and 10 ft. in height. The usual burner, cold-blast, hot-blast, and flue openings were provided; the stoves were heated with natural gas.

Hot-blast mains connecting the stoves were 6-in. iron pipe lined with 4-in. fireclay flue liners, and covered with heavy asbestos insulation. Connections were provided between this pipe and the cold-air supply so that the temperature of the air entering the furnace was always under control. A bypass was provided at one point so that the air could be sent through a rotary gas meter at intervals; the metering was done cold and corrections made to temperature and pressure.


The feeder used was a modified Dunn pulverized coal feeder. Certain parts, originally made of brass, were later replaced by high-speed, hardened, tool-steel parts to avoid excessive wear from abrasion by the ore. A screw conveyor carried the; pulverized ore from the bottom of a hopper to a pipe in which a suction was produced by a compressed-air jet issuing from a 1/8-in. or 3/16-in. nozzle past its lower end. The ore was carried, by an expanding compressed-air current, toward the mixer and injector pipe. The connection into the mixer pipe was made tangentially and also inclined upwards so that the ore and air mixture met the hot air from the stoves in a rising spiral and thus thoroughly mixed the hot air and ore before they entered the combustion tube.

Approximately 10 cu. ft. of free air compressed to 60 lb. per sq. in. were necessary per pound of ore to carry the ore from in front of the 1/8-in. or 3/16-in. nozzle into the mixer. At times, even this amount was insufficient to prevent stoppage of the pipe, so the amount of air was increased at intervals to remove any accumulations by the arrangement shown in Fig. 5.

Seven thermocouples were placed in the furnace: four being placed in the combustion tube as shown, one in the mixer just below the base of the injecting nozzle, one in the downtake at the exit to the flue, and one in the downtake 20 ft. from the base plate. Lead wires were run to a central galvanometer station where readings were taken throughout each test.

General samples from the products accumulated during the run showed these sulfur contents:

This run began 1 hr. before that indicated (9:55) but conditions did not permit sampling, and adjustments caused unreliable results previous to those shown.

Total ore fed during the entire run, 2080 lb.; of this 160 lb. were fed before beginning the record.


Openings for oil burners were provided at intervals of 8 ft. throughout the height of the furnace. These openings were placed tangent to the inside circumference of the downtake. The burners were for the purpose of bringing the furnace to the operating temperature and were not used after such a temperature was reached and feeding of ore had begun.

The products of combustion of the oil heating were taken off at the top until the burners were going well, when the top was closed and the draft directed through the flues near the bottom of the stack. After a temperature of 800° to 1000° C. in the furnace had been reached, the burners were shut off, the burner openings closed, and the ore and air feed started. Adjustments of ore and air feed were then made as indicated necessary by the SO2 content of the gas, temperature conditions, and quality of roasted product. The furnace responded readily to such adjustments, but the time permitted for the test was not sufficient to permit adjustments to the best results possible.


In order to increase the capacity, the 9-in. tube was replaced by a 12-in. This tube was of specially made tile with 1½-in. walls, which restricted the area of the downtake more than was desirable and probably resulted in slightly poorer results than could have been obtained if the diameter of the furnace had been increased in proportion to the increase in the diameter of the combustion tube. Such a change, of course, was impossible without completely rebuilding the furnace.

Table 1 gives the results of a test made with the 12-in. tube. The temperature at the top of the furnace increased beyond that which was desirable, so in the last run cold air was admitted at 20 ft. from the base plate, thus providing for a complete control of the temperature and preventing the hot top. When cold air was thus admitted, it was necessary to reduce the amount of air fed with the ore, which meant a lowering of the velocity in the combustion tube and resulted in an increase in the bottom temperature. This increased bottom temperature, in turn, permitted the reduction in the temperature of the air coming from the stoves.

It was thought that accretions would form in the combustion tube during the roasting but, with one exception, no trouble was encountered from this source, even though the temperatures at times were allowed to go beyond those that would be permitted in practice. Some little difficulty developed at points near the oil burners, as in preheating the temperature in these regions was necessarily higher than the intermediate zones. The temperature of the various zones, however, were quickly equalized and no trouble was encountered during the first few tests in the furnace, the combustion tube being clean at the end of the test; in later tests, however, after some dust had collected in crevices of the tube, the overheating at the burners caused some slagging between the dust and fireclay, thus developing starting points for accretions. It is not thought that this would cause any trouble in practice for, after the furnace is in operation, it will continue so for a considerable time. In this experimental work, the operations were necessarily short and at infrequent intervals, which meant repeated heating of the furnace with consequent trouble from slagging.

The gases and fine dust were taken from the furnace into the flues through four openings equally spaced around the base of the stack. The flue led first into a settling box, in which the rate of flow of the gases was reduced and screens prevented the channeling of the currents. From this settler, the bases passed through a small Cottrell precipitator for final cleaning. The cleaning was very effective, though this end of the operation was given but little attention. The Cottrell equipment contained four 10-in. tubes 16-ft. long which were thoroughly grounded. In the center of these 10-in. pipes, and connected with the 50,000-volt line, were suspended, on insulators, ½-in. pipes fitted with four knife edges at right angles to one another. The 400-volt a.c. current was stepped up to 50,000 volts and rectified by a Kenetron.

The flotation concentrate used in the furnace was a complex sulfide containing 31.4 per cent, sulfur, 44.3 per cent, zinc, 11.6 per cent. iron. The results of the roasting of this material and conditions existing during the roast are given in Table 1. These show a minimum sulfur content in the roasted product of a 2.2 per cent, and 3.6 per cent, for an afternoon period after the furnace had been adjusted to best running conditions. Other tests gave continuous results of slightly over 2 per cent, sulfur.

Tables 2, 3, and 4 show the forms in which the sulfur existed in the roasted product. The particular point of interest in these results is that the sulfur existed largely as sulfate and the quantity as sulfide was small. Table 2 shows the per cent, of SO2 in the gas corresponding to the roasted ore samples. It is evident that the concentration of gas had no effect in preventing the completeness of the roast; i. e., the sulfur is not higher in roasted products produced in higher SO2 concentration.

Note.—Decreased sulfur content of calcines in the afternoon results from the higher furnace temperatures, as shown by Table 1.

Note.—During this run the SO2 content of the gas in the furnace was fairly constant at 9 to 9.5 per cent, during the greater part of the run.

Hopper sample is calcine collected from the hoppers of the furnace; flue sample is calcine cleaned from the horizontal flue leading to the settler; settler sample is calcine taken from the settler; precipitate or sample is calcine taken from the Cottrell precipitator. The coarsest material was from the hopper and the finest from the precipitator.

Tables 3 and 4 show that the sulfur exists in the roasted product as sulfide and sulfate, and that the sulfate sulfur is predominant in the fine ore. This is shown not only by the general samples, of which the hopper samples are the coarsest and the precipitator the finest, but also by the screen analysis of the roasted products. High sulfide sulfur is present only in coarse material, which in total amount is insignificant in flotation concentrate. The total sulfur is highest, in general, in the finest product, but consists largely of sulfate sulfur. Work that cannot be detailed here reveals that this sulfate sulfur is not the result of an incomplete roast m the furnace, but results from a resulfatization of the roasted ore. This resulfatization is due to relation between partial pressures of SO2 and SO3 and the temperature existing in the lower part of the furnace, in the settler, and in the precipitator. Generally, the partial pressure of the sulfur gases is such that no resulfatization of the calcine can take place if the furnace temperature is between 900° and 1000° C. But when this temperature falls to near 800° and below at the flues, the rapid resulfatization of the ore commences. For example, it will be noted that in Tables 3 and 4, the minus 200-mesh material of the hopper samples, i. e., material separated from the furnace at high temperature and immediately cooled, is much lower in total sulfur, as well as sulfate sulfur, than material of the same size from the flue, settler, or precipitator samples, which were subject to the lowering critical temperatures in the presence of sulfating gas. When the aim of the roasting is a practically complete desulfurization, the furnace must be run hot and the calcines and gas rapidly cooled, or separated from each other as soon as they leave the furnace. If the roasting is carried out as a preliminary for sulfuric-acid leaching, as in the electrolytic-zinc process, the sulfating action of the furnace may be used to any desirable degree up to its maximum.

The Proposed East St. Louis Furnace

An increase in height was provided for in the design of the 12-ton furnace it was planned to build at East St. Louis, but which was not completed because of the conditions brought about by the war. Fig. 6 shows that the combustion tube is set to one side instead of concentrically with the downtake. With the concentric combustion tube, it was necessary to place the furnace on posts to provide space for feed and air connections and for repairs. It was also difficult to make pyrometer and air connections from the outside to the combustion tube, at intervals throughout the height, and to provide support for the combustion tube. To overcome these objections, the side combustion tube was designed.

Air and Feed Arrangement

Provision was made for bringing all the air for the roasting from the stoves through the mixer, as in the smaller furnace, or for splitting the current and having a part of the air take this path and the remainder enter the combustion tube near the top of the injector nozzle. Pipe connections to hot- and cold-air sources were placed at intervals along the combustion tube, thus providing for an absolute control of the

temperature in this tube; Fig. 7 illustrates this arrangement of the ore and air feed connections.


As in the smaller furnace, a screw conveyor takes the ore from the bottom of the feed hopper to a pipe immediately above a compressed-air

injector. The injector was changed considerably from that of the smaller feeder, in order to reduce the amount of compressed air necessary for injection purposes. As shown, the compressed-air nozzle blows directly
through a throat, or Eureka siphon, carrying the ore with it and delivering it into a current of low-pressure air coming from hot or cold sources, or both, by which it is carried forward into the mixer.

Experiments indicate that not to exceed 20 cu. ft. of free air compressed to 30 lb. will be necessary for injecting 1 lb. of ore as against 10 cu. ft. of free air compressed to 60 lb. in the feeder used on the small furnace.

The delivery pipe from the feeder was to be connected to the mixer pipe by the tangential and rising connection that was used on the smaller furnace. The same principle was also to be used in connecting the excess-air pipe into the combustion tube.


The bottom, or base, plate of the combustion tube was to be raised about 10 ft. from the floor line, to provide for feeder and air connections. The space opposite these connections in the base of the downtake was to be used for the settling chambers for the roasted ore. Accordingly provision was made for the removal of the coarse product from above a bulkhead in the downtake at a point opposite the injector nozzle. The fines and gas were to be taken off, somewhat above this, through three radially placed flues, which flues in turn entered tangentially a settling chamber placed beneath the bulkhead just mentioned. This method of entry was to provide a circular or centrifugal motion for settling the dust. The gases and remaining dust were to be taken off through a flue passing upward and outward through the bulkhead, thence to the Cottrell precipitator.

Heating the Air

There is insufficient data as to how much of the air must be heated and to what temperature heating should be carried for the most efficient operation of the furnace. The effect of heated air on the behavior of the furnace and the ignition has been discussed and it is obvious that one of four methods must be used.

  1. Heating all the air and passing all of it through the mixer tube and into the furnace with the ore; this will require large stove construction and will not allow air changes in various parts of the furnace, as will be desirable for proper control.
  2. Heating all the air and passing only a part of it into the furnace through the mixer, the remainder to be introduced into the combustion tube at various points; this method is not desirable as the heat is needed in the bottom of the furnace for. heating the ore and bringing about early ignition.
  3. Heating none of the air; indications are that this will not be possible, as already explained.
  4. Heating a part of the air (say one-fifth), which will be passed into the furnace with the ore through the mixer, the remainder of the air to be admitted cold at points throughout the height of the furnace; this method will make all the stove heat available for ignition and for holding down the combustion zone, and will permit the use of cold air for combustion and temperature control of the top of the furnace. Of course, the stoves should be of ample capacity and of such construction as to permit changing the temperature and amount of air going into any part of the furnace as may be found necessary.

Under the first condition, assuming that 300° C. is necessary for mixer temperature, the air in a 50-ton furnace will have to be heated to 350°
in order to heat the ore, if it is being fed cold. If the ore is taken direct
from the dryer, this temperature may be reduced somewhat.

The authors calculate that three stoves of the following dimensions will readily take care of a 50-ton furnace: 1060 sq. ft. heating surface, 246 flues, 4 by 4 in. in cross-section, 14 ft. long, and 3¼-in. walls; total height allowing for gas and air ports, foundation and dome about 25 ft.; diameter about 12 ft.;.cross-section of combustion chamber 12 sq. ft.

Under the fourth condition, which seems the most rational, such stoves would be more than ample, but not necessarily inefficient, and would provide for emergencies where the heating of all the air was necessary. The ratio of fuel necessary for heating all the air to 250° to that necessary for heating one-fifth to 600° is about as 2.5 is to 1, which is quite a saving in favor of the fractional heating.

Drying the Ore

It is obvious that the feed for this type of furnace must be dry, or it will not pass the feeder and injector. It is likely that most of the concentrate delivered at the roasting plant would have to be dried, so it was planned to do this drying with the waste heat in the gases leaving the furnace, by circulating them beneath the hearth of an ordinary kiln dryer.

That there is sufficient waste heat available for this purpose is readily shown. In the experimental furnace, the gas left the furnace at 800° C. under conditions of best roasting. The volume of this gas (measured under standard conditions) was approximately 50 cu. ft. per lb. of ore roasted and was made up of about 8 per cent, sulfur dioxide; 10 per cent, oxygen, and 82 per cent, nitrogen.

Assuming that these conditions are obtained in a 50-ton furnace:

50 tons per day = 69.5 lb. or 31.5 kg. per min.
69.5 lb. will produce 3475 cu. ft. per min. of gas.
8.0 per cent, of 3475 = 278 cu. ft. SO2 = 22.7 kg. per min.
10.0 per cent, of 3475 = 347.5 cu. ft. O2 = 14.1 kg. per min.
82.0 per cent, of 3475 = 2849.5 cu. ft. N2 = 101.7 kg. per min.

Further, assuming that the gas is lowered in temperature by 400° in passing the drier, the heat available there is:

  1. Sulfur dioxide:
    Heat capacity at 800° C. is 1 X 800 (0.125 + 0.0001 X 800) = 164; 164 X 22.7 = 3722.8 kg.-cal.
    Heat capacity at 400° C.
    66 X 22.7 = 1498.2 kg.-cal.
    Heat available from SO2 = 2224.6 kg.-cal.
  2. Oxygen:
    Heat capacity at 800° C. per kg. = 181 cal.
    181 X 14.1 = 2552.1 kg.-cal.
    Heat capacity at 400° C. per kg. = 88 cal.
    88 X 14.1 = 1240.8 kg.-cal.
    Heat available from oxygen = 1311.3 kg.-cal.
  3. Nitrogen:
    Heat capacity at 800° C. per kg. = 207 cal.
    207 X 101.7 = 21051.9 kg.-cal.
    Heat capacity at 400° C. per kg. = 100 cal.
    100 X 101.7 = 10170.0 kg.-cal.
    Heat available from N2 = 10881.9 kg.-cal.
    Total heat available per minute;
    2224.6 + 1311.3 + 10881.9 = 14417.8 kg.-cal.

Assuming 20 per cent, moisture in the ore, it will require 39.3 kg. of moist ore to provide the 31.5 kg. of dry ore. This means that 7.8 kg. of water must be evaporated each minute, and 39.5 kg. of ore must be raised to 200° C.

It is, therefore, evident that the necessary heat is available for drying and heating the ore. If this heated ore is fed without cooling to the furnace, the temperature of the air from the stoves may be reduced somewhat from that necessary when cold ore is fed. As this amount of waste heat was actually delivered by the small experimental furnace, it is clear that enough heat would be available for completing the roast and maintaining furnace temperatures. As this was possible with a small furnace with a high radiation factor, it would be more readily accomplished in a larger well-insulated furnace.

The dried ore from the dryer will contain some lumps, formed in the drying, and some coarse particles that will not work well in the feeder; it was planned, therefore, to interpose a vibrating screen between the dryer and the feed hopper to remove these, insuring a satisfactory feed at all times.


The capacity of these furnaces will vary as the square of the diameter of the combustion tube. This is substantiated in the results from the 9-in. and 12-in. combustion tubes used in the experimental furnace, which had capacities of 2.0 and 3.6 tons per 24 hr., respectively. Assuming similar conditions as to ratio of air and ore, and the velocity in the combustion tube, the diameter necessary in the tube of a 50-ton furnace will be 45 in. It may be that for a furnace of this capacity it would be better to use two tubes having a combined area equal to 45 in. diameter, i. e., 32 in. diameter each. Such an arrangement would provide for repairs in feeders, nozzles, and injectors on a part of the furnace without complete shutting down the furnace, with attendant cooling.