TRANSACTIONS OF THE AMERICAN INSTITUTE OF MINING AND METALLUR- GICAL ENGINEERS [SUBJECT TO REVISION]

. . NO. 1067. I s s u ~ n WITH MINING A N D METALLURGY, FEBRU~RY, 1921

DISCUSSION O F THIS PAPER IS INVITED. It should preferably be presented in person a t t he New York Meeting, February, 1921, wheri an abstract of the paper will be read. If this is impoasi- ble, discussion in writing may be sent t o t he Editor, American Inst~tute of Mining and I\letallurgical Engineers, 29 West 39th Street, New York. N. Y.,, for presentation by the Secretary or other representa- tive of its author. Unless special arrangement 18 made, the diacuss~on of this paper will close Apr. I, 1921. Any discussion offered thereafter should preferably be in the form of a new paper.

Manufacture of Ferromanganese in the Electric Furnace*

BY ROBERT M. KEENEY,~ E. MET., GOLDEN, COLO., AND JAY LONERGAN,$ E. M., WENATCHEE, WASH.

(New York Meeting. February. 1921)

THE electric smelting of manganese'ore and the production of ferro- ' manganese did not exist as an industry, in the United States or elsewhere, previous to the outbreak of war in 1914. Ferromanganese had byen ,

produced, electrically, during times of high prices by two companies a t Niagara Falls and by several European companies, but the production was spasmodic, and none was electrically produced when ferromanganese sold a t from $35 to $40 per ton. Thus the electric smelting of manganese ore must be cbnsidered as a development caused by the high prices prevailing during the years 1916, 1917 and 1918, and the necessity of smelting domestic ore in order to fill the ferromanganese requirements of the steel industry of the United States, which increased from 189,088 tons in 1914 to 356,356 tons in 1918. The electric furnace is well adapted to smelting domestic ore because it can be located in small units near the source of ore supply. In 1918, the ferromanganese production of the United States was 333,027 tons, of which about 7 per cent., or 23,000 tons, was produced in the electric furnace. The electric furnace not only assisted materially in the development of a domestic-ore supply but made a considerable saving of coke, as i t requires about l>i tons of bituminous or lignite coal per ton of ferromanganese compared with a blast-furnace consumption of 2.9 tons of coke per ton of ferromanganese.

Some of the plants in the United States that smelted manganese ore in the electric furnace during the war and in 1920 are listed in Table 1. From early in 1919 to about February, 1920, little ferromanganese was produced in the electric furnace. During this period the price of ferro- manganese (78 to 82 per cent. manganese) was about $100 per ton. During the first half of 1920, several plants resumed operation when the price of ferromanganese (76 to 80 per cent. grade) gradually increased to $200 per ton, delivered.

* A contribution from the Department of Metallurgical Research, Colorado School of Mmes, Golden, Colo.

t Director of RiIetallurgical Research, Colorado School of Mines. f Chief Engineer, Royal Development Co.

--

Copyright, 1921, by the American Institute o j Mzning and Metallu~gical Engineers, Inc.

2 BlAKUFACTURE OF FERROMANGANESE IN THE ELECTRIC FURNACE

TABLE 1.-Electric Ferromanganese Plants in United States

Company 1 1 Capacity, 1 Number / Size of

I Location of Furnace. , Furnaces ; Kv-a.

-- ..... . . . . . . . . . . . . . . . . . . Total. 1 58,000 1 33 1

...... h c o n d a Copper Min. Co. / G r e a t F a 11 s , i 25,000 5

With present coke prices the electric furnace can probably manu- facture ferromanganese as cheaply as the merchant blrtst furnace, because the fuel cost per ton of metal is about the same as the power cost per ton of metal. During September, the price of ferromanganese dropped to $175 per gross ton delivered, due to overproduction and importations from England; in November, the price was $135 per gross ton delivered.

,

As in electric smelting of iron ore, the process consists essentially of the substitution of electric heating for combustion heating, with the use of carbon as a reducing agent and limestone or lime as a flux of the silica in the ore. The fundamental reactions are:

' Mont.

MnOz + 2C = Mn + 2C0 MnC03 + C = Mn + CO2 + CO

MnC03 + 2C = Mn + 3C0 Fez03 + 3C = 2Fe + 3C0

FeO + C = Fe + CO . 2SiOZ + 3Ca0 = 3Ca0.2Si02

2Si02 + 3CaC03 = 3Ca0.2Si02 + 3COz

with excess of carbon there is a tendency to form calcium carbide

Bilrowe Alloys Co.. ............. Tacoma, Wash. ' 6 Ferro Alloy Co.. . . . . . . . . . . . . . . .

Imn Mountain Alloy Co.. .......

Noble Electric Steel Co.. . . . . . . . . Pacific Electro Metals Co.. . . . . . .

Pittsburgh Metallurgical Co.. ....

Southern Manganese Corp.. . . . . . Western Reduction Co. . . . . . . . . . .

Utah Junction, 1 2,% 1 3 Colo.

Utah Junction, Colo.

Heroult, Cali. B a y P o i n t , Cali.

Montour Junc- tion, Pa.

Anniston, Ala. Portland, Ore.

3,000 ' 2

4,500 3,000

1,500

15,000 1,200

5 1

1

8 2

ROBERT M. KEENEY AND JAY LONERGAN 3

which results also in loss of manganese in the slag because of lack of lime to combine with the silica.

3Ca0.3SiO; + MnO = Mn0.3Ca0.3Si02

Excess lime results in the formation of an air-slacking slag due to the formation of the orthosilicate 2CaO.Si02.

SiOz + 2 ~ a 0 = 2Ca0.Si02

The chemical .reactions involved are about the same as in blast- furnace smelting, except that, due to the high temperature a t the endCo_f an electrode in an electric furnace, there is a tendency to form calcium carbide if excess lime or carbon is present. This is due to poor mixing of the charge if lime and coal alone are in contact a t the end of the electrode. On the general grade of manganese ore smelted during the war, the total loss, slag loss, and stack loss seem to be about the same in both processes. In such a comparison, the electric furnace is a t a disadvantage as, from all evidence available, the electric-furnace charges evidently carried less manganese and more silica. With this lower grade ore, the electric furnace accomplished practically the same results as the blast fvhnace smelting higher grade ore.

P. H. Royster' of the U. S. Bureau of Mines, who studied the opera- tion of eleven ferromanganese blast furnaces, gives the results of forty experimental periods of ten days each, the average of which is shown in Table 2. Similar results have been compiled from three months opera- tion of an 1100-kw. electric furnace. The weights given in Table 2 for the electric furnace are determined by actual weighing and all figures are based on metal cleaned and sorted.

The losses in both furnaces are the same, though the electric furnace was smelting an ore containing 34.8 per cent. Mn and 13.2 per cent. SiOz and the blast furnace was smelting ore containing 40.33 per cent. Mn, and 8.6 per cent. SiOz. The electric-furnace slag was less basic than the blast-furnace slag.

he fuel in the blast-furnace char& is thecause of the slag fall from the blast furnace being greater than that from the electric furnace. There is a possibility that the difference in slag fall is caused by the fact that most of the blast-furnace slag was estimated by volume. As is to be expected, the most niarked difference in the two processes is in the carbon consumed, the blast furnace requiring 5323 lb. (2414.5 kg.) per I

1 P. H. Royster: Production of Ferromanganese in the Blast Furnace. Trans. (1920) 62, 18.

4 MASUFACTURE OF FERROMANGANESE IS THE ELECTRIC FURSACE

TABLE 2.-Eldctric Furnaccz versus Blast Furnace

Hlast-furnace figures do not check with electric-furnace figures because of method used to obtain them.

SOTE.-Unless otherwise specified, the gross ton of 2240 Ib. is used in this paper.

Blast Furnace 1 Electric Furnace I - . . . . - . -

Ore per ton metal. pound.. . . . . . . . . . . . . . . . . 5992.0 6544.0 Coke per ton met,al, pound. . . . . . . . . . . . . . . . . 6326.0

ton of ferror~langanese and the electric furnace 1273 lb. (577.4 kg.). On the whole, the metallurgical results obtained by the two methods are the same, except that the electric furnace will apparently smelt an ore containing 5.5 per cent. less manganese and 4.6 per cent. more silica with the same recovery as the blast furnace.

Coal per ton metal, pound. . . . . . . . . . . . . . . . . 200.0 Limestone per ton metal, pound. . . . . . . . . . . . . 2349.0

. . . . . . . . . . . . . . . . Lime per ton metal, pound. Ore analyses:

bln,percent . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40.33 SO2, per cent.. . . . . . . . . . . . . . . . . . . . . . . . . . 8.60

Slag, per ton metal, pound.. . . . . . . . . . . . . . . . . 3196.0 . . . . . . . . . . . . . . . . . . . Metal, per cent. Mn. . . . 74.9

Metal, per cent. Si.. . . . . . . . . . . . . . . . . . . . . . . . 1.15 Slag analysis :

. . . . . . . . . . . . . . . . . . . . . . . . CaO, per cent. hlgO,percent . , 41.75 . . . . . . . . . . . . . . . . . . . . . . . Al2O9, per cent. . . . . . . . . . . . . . . . . . . . . . . . . . . 14.0 SiO-? percent . . . . . . . . . . . . . . . . . . . . . . . . . . 28.1 Aln, percent . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.6

Carbon per ton metal, pound,. . . . . . . . . . . . . . 5323.0 Mn, charged, pound. . . . . . . . . . . . . . . . . . . . . . 2382.0a Mn, per cent. t,o metal . . . . . . . . . . . . . . . . . . . . . 72.0 Mn,percent tos lag . . . . . . . . . . . . . . . . . . . . . . . 14.7 bln, per cent. lost in stack.. ... .,: . . . . . . . . . . . 12.8

The electric furnaces were installed in a wooden frame building with corrugated iron siding and roofing. This const.ruction was rendered necessary by war conditions and caused the furnaces to be placed farther from the transformers than would have been necessary with fireproof construction. The furnace installation consisted of one 1200-kv-a. and one 1800-kv-a. three-phase furnace. A flow sheet of materials is shown in Fig. 1.

I n this plant ore was usually unloaded from railroad cars into 1-ton mine cars, which ran over outside storage bins, Fig. 2, each capable of holding 50 tons of ore, and arranged for the installation of feeder-belt conveyors and trippers in the future. From the storage bins, which had

2839.0 (lignite) 1063.0

253.0

34.8 13.2

2830.0 73.6 2 . 9

37" ) 40.2 2 .4 7 . 9

28 .6 12.4

1273.0 2291 0

72.0 15 .0 13.0 (by difference) '

ROBERT M. KEENEY .AND JAY LONERGAN 5

flat bottoms, the ore was shoveled or wheeled to an 18-in. belt conveyor running the length of the bin system in a concrete-lined pit. Shallow hoppers fed the conveyor.in front of each bin. The ore-bin conveyor discharged on to an 18-in. conveyor running a t right angles to it, and carried the ore u p a slight incline discharging directly into a 9 by 1'5-in. (22.8 by 38.1 cm.) Blake crusher, which discharged the ore into an ele- vator pit. The elevator was inclined from the vertical about 15", and had buckets 6 by 10 in. on a belt 12'in. wide. The elevator discharged

R R Cars

Mine Car I

on to a 48-in. Snyder type sampler, which cut out one-tenth. The reject from the sampler was discharged on to a 14-in. distributing conveyor and tripper running at,right angles to the elevator and over the furnace bins. The sample passed through a short conical mixing trommel and then to a 36-in. Snyder type saxppler cutting one-tenth. The reject was discharged on to the distributing conveyor, while the sample fell to the floor below, "where i t was quartered. The reject from the sampler floor was shoveled into the elevator and discharged on to the distributing conveyor. Cars of ore that had been sampled in transit were not sampled

a t the plant. Cars could also be unloaded directly on to the crusher conveyor. Cars that were to remain in storage were sampled by shovel, taking every tenth shovel.

FIG. 2.-STORAGE BINS.

Furnace

The furnace consisted of a rectangular steel shell 18 ft. long, 8 ft. wide, and 7 ft. deep made of >$-in. plate, riveted and stiffened with

ROBERT M. KEENEY AND J A Y LONERGAN 7

channel iron, and rested upon four concrete piers. The three electrodes were hung from steel I-beams' supported by the wooden trusses of the building. The furnace was operated with electrodes 17 in. (43 cm.) in diameter by 72 in. (182 cm.) long; some of them were threaded for con- tinuous feeding and some were "butt" electrodes. When the furnace was started, it was impossible to obtain 17-in. electrodes, so 16-in., 14- ' in., and 8-in. were used. The load on the furnace was regulated by

'

controlling the electrodes by hand. The sides and bottom of the furnace were lined with 8 in. (20.3 cm.) of firebrick, and this was lined with a mixture of calcined white California magnesite and hot pitch, which was tamped in place. There was no roof to the furnace, the charge level being left as high as possible; Fig. 3 is a section through the middle electrode and the tap hole. Metal and slag were tapped into pot cars through the same tap hole. The furnace was surrounded by a charge floor, which was level with the top, and was charged from bins that dis- charged on to this floor.

Transformers

Power was supplied to the furnace by three 400-kv-a., single-phase, Allis-Chalmers, oil-filled, self-cooled transformers of the following speci- fications: primary voltage, 13,200; secondary voltage, 75/3749; inherent reactance, 10 per cent., 2-5 per cent., primary taps.

The transformers were connected with coils in series delta-delta, Aith the primary sides on the 10 per cent. taps because of low line voltage, resulting in a voltage of 75 volts a t the transformers. The secondary side was brought out of the transformer case as bushars. These transformers operated very satisfactorily, but on the hottest days in summer, with a load of 1100 kw., 85 per cent. power factor, there was a temperature rise of about 50' C., so air was blown against the case to prevent further rise.

Electrode Holders and Secondary B u s System ,

The electrode holders first used and the method of supporting the holders were based on the operation of furnaces taking electrodes up to 12 in. (30.5 cm.) in diameter, as these methods were satisfactory for the smaller furnaces. Although holder and cable troubles developed, the furnace was always operated a t a profit and a t over 78 per cent. average .load factor. The holder on a smaller furnace, besides being subjected to less strain because of the smaller electrode, was not subjected to as high temperatures, because of the smaller volume of hot gases that passed from the furnace.

The requirements for a satisfactory holder are: Sufficient contact area between copper and carbon; mechanical strength; electrical conductivity; sufficient water cooling to prevent slipping and corrosion of the electrode

in the contact area; absence of copper to copper connection over the furnace or where there is exposure to hot gas; absence of small bolts or small threads which are not water cooled.

In the first holder used, shown in Fig. 4, the electrode was held by two levers, which forced the clamps against the electrode. The objec-

FIG. 4.-16-IN. LEVER-TYPE HOLDER, FURNACE NO. 1.

FIG. 5.-16-IN. BOLT TYPE HOLDER, FURNACE N O . 1.

tions to this holder were: The lever arms tended to swing together, causing short circuits; the pressure of the levers was so great that the holders were bent; there was short circuiting through the lever bolts between the two clamps; the holder was of too light construction and cracked after continued use; there was not sufficient water cooling; and as the connection beween the flexible cables and the holder was made

ROBERT M. KEENEY AND JAY LONERGAN

, by bolting the cables to a short bus, which in turn was bolted to the holder. both connections being enveloped in the furnace gases, clean contacts could not be maintained because of shearing action of the busbar 'and cables a t their contacts. These holders were installed with the two halves of each holder parallel to the long axis of the furnace, but after a

FIG. 6.-16-IN. HINGE-TYPE' HOLDER, FURNACE N O . 1.

short period of operation the levers were replaced by bolts. This was not much of an improvement because of the shearing action a t the con- tact between busbar and holder.

The second holder, shown in Fig. 5: differed from the first in that bolts . and nuts were used instead of levers, and the connection of the flexibles to the busbar was placed farther from the hot part of the furnace. I t

helped matters considerably but did not remove the difficulty of poor "

contact between busbar and holder, and whenever an electrode was changed or slipped it was generally necessary to cut the bolts. A con- siderably higher load factor was maintained with this holder, but con- tinual repair work was required.

A third holder, shown in Fig. 6, met all the requirements, except that i t was not strong enough on the front side, which later was strengthened by casting i t with stiffening ribs. The holder was of the hinge type and consisted of two parts cast of 95 per cent. copper, with no busbar con- nection.. The flexible connection to the holder arms was a t the edge of the furnace shell. There were no bolts or levers, but three simple pins. I t had a large water-cooled contact area. The contact area between the copper and the carbon was based on a load of 15 to 20 amp. per sq. in. of

E C W E C W E C W

00 000 0 000000000

E = From Easf C = From C e n t e r W =From West

I, ooqo00 C i r c u l a r M i l . Cable

1200 K.W. Furnace

PIG. 7.-CABLE INTERLACING DIAGRAM.

maximum contact area with a perfect electrode. The holder was de- signed so that the area of one half of the holder would carry the complete load if required.

The system of holder suspension from I-beams supported on the roof trusses was unsat.isfactory, although it works all right onsmall electrodes. At first, the holders were not counterbalanced, which resulted in their tilting a t an angle dependent on the center of gravity of the electrode. The steel cable$ supporting the holders ran to sheaves on I-beam trolleys; these trolleys frequently failed to work because of the dust. The holders' and the electrodes were eventually counterbalanced. The absence of a steel hood over the furnace resulted in fires, there being merely the usual foundry ventilator with louvers., An open-top furnace installed in a wooden building should always have a hood, especially when coal is used as a reducing agent.

The low-tension sides of. the transformers were connected in delta.

ROBERT M. K E E N E Y AND JAY LONERGAN 11

From the delta connection over the transformers each phase was con- ducted to busbars at the furnace, as shown in Fig. 3, by ten 1,000,000 c.m. copper cables, with weatherproof insulation, to a phase. Dossert- connected lugs were used on the cable. The current was conducted from this point t o the electrode holders by forty 0000 bare copper cables. The 1,000,000 c.m. cable was supported by insulator racks, and.all phases were interlaced as shown in Fig. 7, up to about 2 ft. .(60.9 cm.) from the point where connection was made to the flexibles. Due to their burning out, the 0000 flexible cables were replaced by copper strips 2 in. wide by x6 in. thick, 7 f t . 3 in. long (50.8 mm. by 1.59 mm. by 2.21m). These gave excellent results, eliminating all lugs in the furnace gases. Except where used over the furnace, the Dossert lugs gave fair satisfaction. After about four months continuous operation, the insulation on the 1,000,000 c.m. cables softened and ran down into the Dossert lugs on the connection a t the delta bus, and caused bad contacts. On furnace No. 2 all .lugs were replaced by welded connections. The loop from the delta connection on the transformers to the end of the electrode on furnace No. 1 consisted of 20 ft. 6 in. 1,000,000 cm. cable, 7 ft. 3 in. copper strips, 3 ft. 3 in. electrode holder, 4 f t . electrode, making a total length of 35 f t . (10.67 m.). The use of 1,000,000 c.m. cable instead of a busbar was fairly satisfactory and gave entire satisfaction on furnace No. 2, where welded lugs were used on the cable. Its main advantage was the ease with which complete interlacing of phases could be obtained. If a busbar had not been so difficult to obtain i t would have heen used.

FURNACE NO. 2, 1800 KV-A.

Furnace

, The second furnace, Figs. 8 and 9, consisted of a rectangular steel shell 18 f t . long, 9 ft. wide, and 7 ft. deep, made of>.i-in. (12.7mm.) plate, riveted and stiffened with channel iron and resting upon four concrete piers. The three electrodes were hung from a steel superstructure, Figs. 3 and 9, that was entirely independent of the building. Over this superstructure ran a hand crane with an electric hoist, which was used for installing electrodes. The furnace was operated with electrodes 24-in (61 cm.) in diameter, threaded for continuous feeding. The load on the furnace was regulated by hand control of the electrodes by chain blocks. The furnace was lined with an 8-in. (20.3 cm:) wall of red brick on the sides and bottom, and inside the brick, a lining was made by tamping in place a mixture of calcined white California magnesit.e and hot pitch, as shown in Fig. 3. There was no roof. Metal and slag were tapped into cars through the same t,ap hole. The furnace was surrounded by a charge floor that was level with the top, and was charged from bins that dis- charged on to the charge floor.

12 MAXUE'ACTURE OF FERROMANGANESE I S THE ELECTRIC FURNACE

Transformers

Power was supplied to the furnacc by three 600-kv-a., single-phase, General Electric, oil-filled, self-cooled, radiator-type t.raasformers of the following specifications: Primary voltage 13,200; secondary voltage 75; inherent reactance about 7 per cent., 3-5 per cent. primary taps.

The transformers were connected delta-delta with the primary sides on the 10 per cent. taps, resulting in a voltage of 75 volts a t the trans- former. The secondary side was brought out of the transformer case as busbars, which projcctetl 12 in, above thc top of the transformer cover. This projection is not enough when threc large single-phase transformers are connected in delta and the poles are interlaced; the bus in this case shoulcl have projected 24 in. If the secondary-delta bus

FIG. 8.-TAI'PING FLOOR, FURNACE KO. 2 , S I I O W I S G POT CAItS AXD STINGER

connection is within 8 in. (20.3 cm.) of the cover, heating results. If the transformer busbar is only 12 in. (30.5 cm.) long, either i t must be built up 12 in. more, or an unsightly connection used. These transforn~ers *re operated with a load of 1100 to 1150 kw. and were never loaded to full capacity because of the condition of the ferromanganese market.

Electrode Holders and Secondary Bus System

The electrode holders of this furnace and their method of support gave excellcnt results. The electrode was held by a 24-in. hinge holder, Fig. 10, silnilal in design to the third type of holder used on furnace No.1 but of greater strength and lower copper content, 90 per cent. This holder gavc cntire satisfaction, and no trouble was experienced with the

- ROBERT M. KEENEY AND JAY LONERGAN

holder or the flexible connections. The basis of design for current- carrying capacity was the same as for the furnace No.1 holder.

a By use of the 3-ton electric hoist on the hand crane, electrodes could be slipped or replaced in 20 min. as compared with 1 hr. for changing electrodes on furnace No. 1, with hand-operated chain blocks and trolley. The frameworli was covered with a two-stack steel hood, which removed all furnace gases and reduced the danger of fire to a minimum.

FIG. 9.-FRONT VIEW, FURNACE NO. 2, FROM CHARGE FLOOR. FURNACE IS EMPTY.

The low-tension sides, of the transformers were connected in delta. From the kdelta connection over the transformers, each phase was conducted to a busbar a t the furnace, as shown in Fig. 3, by twenty 1,000,000 cm. copper cables with weatherproof insulation. Welded connections'were used on both ends of each length of cable. The current was conducted from the busbar to the electrode holders by copper strips 2 in. (5 cm.) wide by M6 in. (1.6 mm.) thick by 10 ft. (3.05 m.) long. The 1,000,000 cm. cable was supported by insulator racks, and all phases were interlaced up to about 2 ft . (61 cm.) from the point where connec- tion was made to the flexibles, Fig. 7. Neither the welded lugs, the cable,

nor the strips gave any trouble during the operation of the furnace. One side of the loop from the delta connection on the transformer to the end of the electrode on furnace No. 2 consisted of 1 9 ft. of 1,000,000 cm. cable, 10 f t ; of copper strips, a 445-ft. electrode holder, and a 4-ft. elec- trode, giving a total length of 37 ft. 6 in. (11.4 m.).

Fro. 10.-24-IN. HINGE-TYPE HOLDER, FURNACE NO. 2.

MANGANESE ORES

The ores used in the manufacture of ferromanganese, or furnace' ores5 are divided into two classes-oxide and carbonate. The first class is made up of pyrolusite, MnOz, manganite, MnO(OH), hausmannite, MnsOd, psilomelane, MnOz + (HzO.K20.CaO), wad, rhodonite, MnSiOa, and braunite, Mnz03. Domestic ores are a mixture of these oxides and a s

ROBERT M. KEENEY AND JAY LONERGAN 15

such cannot be readily distinguished from each other. For practical purposes it is necessary to call them pyrolusite. Rhodonite is an un- desirable ore that cannot be made into low-silicon ferromanganese. I t is included here because the oxides are usually alterations of it, and i t may be present in the mixture in varying quantities. When present, i t usually passes into the slag without reduction. Rhodochrositc, MnC03, is mined on a large scale in only one locality, Butte, Mont. It has proved a desirable ore for electric smelting when mixed with oxide ore, but when smelted alone some difficulty has been experienced in maintaining a desirable output of metal.

For practical purposes oxide furnace ores are classified into: high- grade ore, which contains 35 per cent., or more, manganese with iron usually below 5 per cent., and low-grade ore, which contains from 10 to 35 per cent. manganese and from 10 to 25 per cent. iron.

The authors have used both steel turnings and ore containing 30 per cent. manganese and 10 per cent. iron for addition of iron to the charge when smelting high-grade ore containing little iron. The use of ore instead of turnings proved more satisfactory, because the furnace seemed to operate more uniformly and give a greater output. The question of the use of iron ore or a manganiferous ore is one of costs.

Analyses of the ores smelted in obtaining the data given in this paper are shown in Table 3. Of these ores, the Ely ore smelted the most rapidly. After the use of steel turnings was abandoned, Leadville ore was charged to supply the iron. The combination of Ely and Leadville ores gave a mixture that smelted more rapidly than any other tried. The slowest smelting ore was the Las Vegas. This f as due to its fineness, as the ore goes to a powder and resembles bog-iron ore, which resulted in furnace blows and a heavy dust loss. Because of the large proportion of coarse

TABLE 3.-Domestic Manganese Ore from Colorado, Utah, and Nevada

Mn, per cent.. . . SiOz, per cent. . . XgO, per cent . . Pe, per cent.. . . . CaO, pcr oent. . . AltOa, per cent. . P, per cent. .. . . . h, p& cent. .. . . . B a q per cent.. . Pb?, per cent.. . Zn, per.cent.. . . . Cu, per cent . . . HaO. per cent:. .

' Cot, per cent . .

Lead- ville. Colo.

33.10 11.55

11.7 1.6

12.92 0.066 0.109

4.7 trace none

12.6

42.12 15.42 trace 2.65 7.94

0.023 0.081

5.65

Ely, Nev.

38.92 13.3 trace 4.26. 0.80 3.96 0.022 0.052

13.2

34.54 34.9 12.8 13.0 0.24 0.65 8.11 5.71 4.23 7.65 3.26 2.31 0.033 0.03 0.73 0.58 2.66 1.92 prwent , present

. Las Vegas, Nev.

W. Tintic, Utah

-

Green River, Utah

Composite Sample Dry Ore to Furnace, 1818

October I Nobember 1 December

material, Leadville ore is fast. smelting. The difficulties of operating this plant were considerably increased by the fact that it was necessary to change the charge almost every day, because the ore was so variable, coming from several properties, and with such a small operation, bedding did not seem practical.

RE~uc1h.a AGENTS AND FLUXES

Several reducing agents have been used in electric smelting of man- ganese ores, and although lignite and bituminous coals were considered the best during the war, the operating results of one producer show that coke can be used satisfactorily. The supposedly essential characteristics of a good reducing agent are a high ratio of fixed carbon to ash, minimum quantity of highly combustible gases, and low electrical conductivity. Table 4 gives the analyses and ratios of carbon to ash of reducing agents available in Colorado.

TABLE 4.-Colorado Reducing Agents

Segundo coke.. . . . . . . . . . . . . . . . . . . . . . . . . 0 . 6 8 , 1.7 Crested Butte, anthracite slack.. . . . . . . . . 0.68 7.82 Crested Butte. anthracite slack.. . . . . . . I , 6.62 Crested Butte, anthracite No. 6 . . . . . . . . . I I 7 .4 Crested Butte, anthracite No. 6 . . . . . . . . 5.87 Leyden nut, lignite coal.. ............... 13.84 36.42 Leyden, mine run, lignite coal. . . . . . . . . . . 13.0 . 36.44 ~Iatrhlem, lignite coal.. . . . . . . . . . . . . . . . . 46.46 Oak Creek, lignite coal . . . . . . . . . . . . . . . . 6.69 1 25.6 Reraind, lignite coal . . . . . . . . . . . . . . . . . . 1.66 41.71 Sopris. bituminous coal.. . . . . . . . . . . . . . . . ; 1.26 / 33.64

Per Cent.

The commonly used fluxes are limestone and lime. Limestone low in phosphorus and magnesium is desirable. On the whole limestone is preferable to lime, as lime is difficult to hold in storage bins because of slacking and expansion. I t slacks so easily that after a short period of storake the CaO content has changed, and the slag analysis varies be- cause of uncertainty as to the exact quantity of CaO being charged.

In ccmmercial operation, a complete charge calculation serves only

Volatile Matter, 1 6 Carbon. 1 Ash,

Per Per c',":t. ; Cent. Cent. ' Cent.

as a guide in starting up the furnace, because coal, limestone, and the iron-bearing material are varied in the quantities experience has shown to be best.in order to produce metal of the manganese and silicon content desired and a slag containing 10 to 15 per cent. manganese. E. 5.

$~t& $h

Bardwel12 has given a typical theoretical charge calculation. -- - - -- - - -- -

E. S. Bardwell: Electric Furnace Smelting of Montana Manganese Ores. Chem. and Met. Eng. (Apr. 14, 1920) 22, 681.

ROBERT M. K E E N E ~ AND JAY LONERGAN 17

For use when the iron in the charge is supplied by a manganese ore comparatively low in manganese and high in iron, some curves have been

Ra+io of Mn +Q Fe l?@q~;red in Ore Chorge

Fe Difference,Percent

. . FIG. CHARGE CALCULATION CURVE, ORE PROPORTIONS:

developed, which aid in quick charge calculation. The curves shown in Fig. 11 provide a quick means of determining the ratio of manganese

to iron required in the charge for production of ferromanganese of a certain grade, considering the approxiinate recovery of manganese being made at the time.

Having determined the ratio of manganese to iron, the proportions of any two ores required to give the desired grade of ferromanganese can be determined from the curves shown in Fig. 12 where S o . 1 ore is a high-grade nlanganese ore with low iron content, and No. 2 ore is a low- manganese high-iron ore. Or the proportions can be found from the following formulas in which the total weight. of the mixture is 1000 lb. (453.6 kg.). Weight of No. 1 ore =

1000 [ R (per cent Fe in No. 2 ore - per cent. Mn in No. 2 ore)] -. .- -. - . - - - Difference in per cent. Mn + R (difference in per cent. Fe)

This formula is developed as follows, letting weight of mixture be 1000 lb. A = high-grade ore, No. 1, in pounds; B = low-grade ore, No. 2, in pounds; x = per cent. manganese in No. 1 ore; y = per cent. iron in No. 1 ore; z = per cent. manganese in No. 2 ore: to = per cent. iron in No. 2 ore;

Total manganese in mixture R = Total iron in mixture

Then

H = Ax + B2 AY +Bw

B = (1000 - A) A x + Bz = RAy + RBw Ax + z (1000 - A) = RAY + Rw(l000 - A) Ax + lOOOz - Az = RAY + 1000Ru~ - ARw Ax - Az - RAv + ARw = lOOORw - lOOOz A(Z - - R~ + RW) = IOOO(RW - 2)

1000 (Rw - z) - - - 1000 (Rw-z) A = x - ~ - R ~ ~ + R w Z - z + R : w - ~ )

1000 [R(per cent. Fe in No. 2 ore - per cent. Mn in S o . 2 ore)] A = - - - -- - - - - - - - - . - - - - --- Difference in per cent. Mn + R(difference in per cent. Fe)

Furnace KO. 1 was operated for about 8 mo. with an average load of 1100 kw. measured on the primary circuit, and a t a secondary voltage of 70 to 75 volts a t the furnace. Except for difficulty with the faulty holder design and the time required for changing electrodes, there was little operating time Iost and the electrical load factor averaged 78 per cent. for the 5 mo. electrodes were available for full monthly operation. The'power factor was 85 per cent.

Furnace No. 2 was operated for about 1 ) i mo. with an average load

ROBERT M. KEENEY AND JAY LONERGAN ' 19

of 1100 kw. measured'on the primary circuit, and a t a furnace voltage of 70 to 75 volts. With the improved electrode holder design and 24-in. electrodes instead of 17-in. electrodes, as on furnace No. 1, a high load factor was obtained. The electrical load factor was 85.1 per cent..,and the furnace operated 96.4 per cent. of the total time; this improvement was due to the better mechanical arrangement for changing electrodes, no holder troubles, and the lower electrode consumption requiring fewer changes of electrodes.

The furnace was charged and tapped every 2 hr. Ore, coal, and limestone were drawn from the bins feeding onto the furnace floor, where they were weighed and spread mixed. All of the charge for a 2-hr. period was charged immediately after tagping, which is contrary to the practice in most plants where the furnaces are charged a t about 15-min. intervals and poked often. The maximum size of the ore was about 244 in. (63.5 mm.) with a large percentage of fine material. Nut lignite coal was used. When anthracite was charged, i t was either slack (mostly dust) or No. 6, which is the %-in. size. The lime rock was crushed to 234 in. Metal and slag were tapped a t 2-hr. intervals through a common tap hole into pot cars. The tapping was done'by driving a bar into the hole and then driving it back out; a t times i t was necessary to use either an electric "stinger," Fig. 8, or oxygen.

DIFFICULTIES OF FURNACE OPERATION

Although the electric smelting of manganese ores is not difficult, from the metallurgical standpoint, in that the grade of the product can

' - be readily controlled, several operating difficulties may arise: Formation of carbides in furnace, blowing, failure of bottom, tapping, failure of electrode holders, and bridging of charge.

Formation of Carbides in Furnace

Formation of carbide in electric smelting of manganese ores for pro- duction of ferromanganese is common and may be expected a t any time. No other metallurgical condition results in greater operating troubles. The presence of carbide is shown by the odor of acetylene when the slag is dropped in water, by crusts rising in the furnace and collecting around and under the electrodes, by-difficulty of tapping, and by either a thick slag or a foamy slag. In placing new plants in operation, the common error has been to attempt to produce a slag low in manganese, 5 per cent., by operating with a considerable excess of reducing agent and a high-lime slag. The combination of considerable excess coal and sufficient lime to give a slag cont'aining much over 40rper cent. CaO invariably results in . the formation of carbide. By operating with such a charge a 5 to 8-

20 MANUFACTURE OF FERROMANGANESE I N T H E ELECTRIC FURNACE

per cent. manganese slag will be produced for several days, but before a week has passed the furnace is usually full of carbides, so that it can be tapped only with great difficulty, and there is a marked reduction in the output of metal. The electrodes tend to climb out of the furnace, be- cause of the accretions of carbide that have formed beneath them, and the metal freezes on the bottom of the furnace. The authors found that continuous charging resulted in a greater tendency to formation of car- bide than intermittent charging, and that carbide formed more readily with anthracite than with bituminous coal, and more readily with bituminous coal than with lignite. When carbide is forming in large quantity in the furnace, there also appears to be considerably greater volatili- zation of manganese. Carbide does not, however, affect the total per- centage of manganese recovered to any extent, because the increased volatilization loss seems to be offset by the lower slag loss. A part id solution of the carbide difficulty was made by the following procedure:

1. Use of lignite instead of anthracite or bituminous coal. 2. Reducing the quantity of coal to slightly above the theoretical

requirement, and reducing the quantity of limestone in the charge, so that the slag contains 12 to 15 per cent. manganese, and about 40 per cent. CaO plus MgO.

3. When carbide forms in spite of the above precautions, the use of a flushing charge containing the usual quantity of all ingredients except coal, which is reduced to one-half the usual quantity.

The operating data compiled in Table 5 show, in a general way, the results of changing from anthracite to lignite, the effect of operation with a lower lime slag, and the effect of flushing charges. So many variables enter into data obtained from plant operation that the cause of the increased production during the last four months of operation cannot be assigned to any one cause, but must be considered as due to the combination.

Blowing

By "blowing" is meant eruptions in the furnace caused by the accumulation of gas in the charge, which escapes under high pressure and burns on the top of the furnace. With an open porous charge, this would not occur to any extent and is not serious in the manufacture of ferro- chrome or ferrosilicon, where coke is generally used as a reducing agent. The blows of a manganese furnace oper;ted with coke should not be . serious, but with coal as a reducing agent, a t times they are very serious. Carbon monoxide from reduction accumulates in pockets with gradually increasing pressure, and as manganese ore is generally fine and dust,y the charge packs, so that the gas can not escape until its pressure forces i t out of the charge. The charge may be blown 25 ft. above the furnace, carrying with i t fine particles of coal, which ignite and cause a mass of

ROBERT M. KEENEY AND JAY LONERGAN 21

flame over the furnace. This may cause the suspension chains of the holders to break, and will eventually cause trouble with the holders if there is any copper-to-copper connection, over the top of the furnace. There is a considerable loss of manganese in dust and volatilization at each blow.

The authors started their operation on the Las Vegas ore, of the analy- sis given in Table 3, with slack Colorado anthracite as the reducing agent. This ore is of the bog type and when crushed goes to a powder that is unusually light and dusty. This ore was smelted during May, June, and July. In combination with the fine anthracite charged during May and June, the furnace was blowing a large part of the time, which together with the formation of carbide resulted in a low output. The unaccounted loss, or dust and volatilization, amounted to an average of 26.9 per cent. of the manganese charged. Considerable of this loss was due to the dusty character of the ore, as slack anthracite was used in January on a coarse ore, for economic reasons, when the unaccounted loss was 12.1 per cent. When nut lignite was used the blows were much less serious. Except for its tendency to form carbide and its high cost, coarse anthra- cite would probably have helped to stop blowing more than the nut lig- nite, which always contains fine material due to slacking before use. This fine coal was screened out on the charge floor.

The authors found that the most satisfactory results, from the view- point of carbide formation, blowing, and labor costs, were obtained by charging the furnace full after a tap and not touching it except in case of a bad blow until after the next tap. Continuous charging and poking tended to aggravate the blowing. This seemed to be due to the fact that the furnace could not get settled in any fixed condition of smelting, and the poking of the charge into a hole, with possibly molten material a t the bottom, always started a blow.

'The fundamental cause of blowing in a ferromanganese furnace is the nature of the charge. Manganese ores contain a large percentage of fine materials; particularly rhodochrosite, which decrepitates and goes to a powder in the furnace, giving off a large volume of COz, which in- creases the blowing. Flame has been observed ) blow to a height of 50 ft. for several minutes from the top of a 3500-kw. furnace, smelting rhodochrosite with semi-bituminous coal as a reducing agent.

Failure of Bottom

To the knowledge of the authors only three materials have been tried for furnace bottoms-carbon, dead-burned or calcined magnesite, and firebrick. The carbon bottom isused at Anniston, Ala., and Bay Point, Calif. The magnesite. bottom was used a t Utah Junction, Colo. At Great Falls, Mont., both carbon and magnesite have been used. I t is reported that ordinary firebrick was used a t Heroult., Calif.

22 MANUFACTURE O F FERROMANGASESE IN THE ELECTRIC FURNACE

One plant reported that the carbon bottoms failed every 2 or 3 mo. until the furnace shell was elevated high enough above the floor to allow air cooling. The magnesite bottoms installed by the authors never failed. One was in continuous use for 8 mo. and the other If5 mo. The first magnesite bottoms installed in another plant were in use for several months and did not fail until the cross-section of the electrodes was reduced one half, so that the full load of the furnace was put through the reduced section, thus overloading them 100 per cent. I t is understood that the firebrick bottom tried a t Heroult, Calif., was satisfactory, but little information regarding it is available.

Theoretically magnesite should make the best bottom. The slag formed in smelting manganese ore must be essentially basic, because if i t is acid manganese acts as a base and substitutes for the deficiency in lime. Also magnesite does not easily form combinations with manganese or carbon. On the other hand carbon combines with manganese to form manganese carbide, with lime to form calcium carbide, and with the oxygen in the ore. The failure of carbon bottoms is probably due to a combination of these reactions, but may be due to careless operation of the furnace and lack of air cooling, factors that would influence the length of life of all linings, if they affect the life of one.

Careless furnace operation has probably ruined many bottoms, particularly in large plants. I n open-top furnaces, there is a tendency for craters to form under each electrode, due probably to the intense heat a t the end of the electrode. These craters fill with molten metal, which remains there. If the crater gets too deep, the bottom will go. When- ever the electrodes are riding low in the furnace and a considerable depth of molten metal is below the level of the tap hole, an attempt should be made to freeze this metal, either by variation of the charge so as to cause the electrodes to rise, or by throwing off the power until it freezes. The authors have had under the electrode pools so deep that only a foot of bottom was left, and have not lost the bottom.

When installing a magnesite lining, only enough pitch should be used to make the magnesite stick together. The bottom of the crucible should be about 1 ft. (30.48 cm.) above the top of the tap hole level, as it will drop about 1 ft. when the pitch burns put. The sides should be built of magnesite only for about 2 ft. above the bottom of the crucible, and capped with firebrick, because the magnesite on the sides tends to cave.

Tapping

Tapping is apt to be difficult when there is considerable carbide in the furnace and when the electrodes are riding high; a t other times the hole can generally be opened without much trouble. A better method of tapping than driving the hole was recently observed at one plant. The

ROBERT M. KEENEY AND JAY LONERGAN _ 23

bar is driven into the hole immediately after tapping and plugging and - is left there until the next tap; then driving i t out usually opens the hole

without much sledging or the use of oxygen. Tapping metal and slag into pots, which are allowed to cool before

dumping, effects considerable saving in labor over tapping on to the floor or pouring from a ladle on to a casting machine; the metal is better in appearance and is more free from slag. The labor saving is in the men required for cleaning the metal. When metal and slag are tapped into a ladle, the slag skimmed, and the metal poured on to a casting machine, there is a high percentage of skulls and dirty slag, which must be concen- trated by wet methods. This is avoided in pot casting, when the slag containing metal is such a small proportion of the total slag tapped '

that i t can be charged back into the furnace daily. I t will average less that 10 per cent. of the weight of uncleaned metal tapped. The increased labor of floor casting or machine casting over pot casting is due to the fact that the pot-cast metal is in large chunks, so that it can be more easily cleaned and sorted; also, no labor is required for wet concentration of skulls and dirty slag. One man can clean and sort about 10 gross tons of pot-cast metal per 8-hr. day.

The pots used for pot tapping were made of cast iron. At first many pots were lost, but during the last 4 mo. of operation the loss was very small. The burning out of pots is due to filling them too full of metal, which results in softening the iron in the pot so that a hole is burned through a t some point. By filling the pot only two-thirds full, this is avoided and the life of the pot is greatly increased. Where a cast-iron ladle is set in a pit under the runner, and the metal and slag fall about . 6 to 8 ft., it is necessary to put a slag bottom in the pot, or i t will burn through. When tapping into pot .cars, the metal fell only about 2 ft. so that only a fireclay wash was necessary.

Failure o j Electrode Holder

Failure of electrode holders is usually due to defective design and to < imperfect electrodes. The design of a holder may be faulty in three

' respects: The holder has a copper-to-copper connection over 'the top of the furnace'; there is not sufficient contact area between copper and carbon; and the design is not sufficiently rugged and bends out of shape under heavy usage or when clamping an undersize electrode.

Improper design has been due to failure to recognize the difference between conditions on the top of the electric manganese furnace and the electric steel or other ferroalloy furnaces. The top of the electric.stee1 furnace is cool, and a holder in which horizontal busbars are bolted to the electrode clamps works satisfactorily. The electrode holder is not con- tinually exposed to the heat and flame that rise above the top of the manganese furnace and envelop the holder. The ferrochrome furnace is

not nearly as hot above its top as the manganese furnace and seldom blows. The electrode holder of the ferromanganese furnace should contain no busbar; instead an arm should be used which is cast with one side of the clamps, so that the holder consists essentially of two pieces, as shown in Pig. 10. The contact area between copper and carbon should be figured on a basis of carrying not to exceed 20 amp. per sq. in. The walls of the holder should be well reinforced with ribs to prevent bending, as much time is lost in slipping electrodes into a warped holder. The walls of the holder around the electrode should be cored and complete water cooling provided. Lack of sufficient water cooling a t the clamps causes slipping of the electrode in the holder and arcing between the holder and the electrode. I t is preferable that the holder be hinged, as it eliminates the use of one bolt or of a pin-and-wedge connection. The authors prefer the pig and wedge on a moderate size furnace as it is simpler and if properly designed seldom gives trouble. If a bolt is used, it should be water cooled or of large dimensions.

Bridging of Charge

Bridging of the charge between the electrodes is caused by the elec- trodes being spaced too far apart. When furnace No. 1 was started the electrodes were spaced on 5-ft. (1.5 m.) centers but the charge bridged SO that metal and slag under the end electrodes could not be tapped. The electrodes were then spaced on 4-ft. centers and later were spaced onj 345 ft. centers, then there was no bridging. The electrodes on furnace No. 2 were spaced on 355-ft. centers when the furnace was started and the charge never bridged.

The results of smelting 6,355,311 lb. (2882.7 T.) of manganese ore in an 1100-kw. furnace of the type described, with the production of 2,083,- 818 lb. (945.2 T.) of ferromanganese, are shown in Tables 5 , 6, and 7. In drawing conclusions from these tables it must be remembered that these are plant operating results with several variable factors. In May, due to shortage of electrodes, four 8-in. diameter electrodes were used per holder instead of 16-in. electrodes; 16-in. electrodes were used in June and July. In August, September, October, November, and December, 14-in. and 17-in. electrodes were used. In December, at about Christ- mas, the operation of furnace No. 1 was stopped, and furnace No. 2 started with the same power input, but with 24-in. electrodes. Some of the 14-in. and 17-in. electrodes were "butt" electrodes,.and some were threaded; the other sizes were threaded. The electrodes were of four brands. During all of May and June and part of July, the required iron

ROBERT M. KEENEY AND JAY LONERGAN 25

was supplied by steel turnings. Late in July, Leadville iron-manganese ore was substituted for the turnings. Anthracite was used entirely in May and for the greater part of June. After that nut lignite was the reducing agent, except in January, 1919, when anthracite was again charged in order to consume the stock on-hand. Limestone was the flux in May, June, December, and January. In the other months both lime- stone and lime were charged. In July, the furnace was shut down seven days because of lack of electrodes, in August three days, and in September seven days. The grade of ore gradually dropped from 39.1 per cent. manganese, in May, to 34.25 per cent., in January.

Table 5 shows the metallurgical results; Table 6 shows the consump- tion of furnace supplies; and Table 7 the power consumption. These tables are to be used as a whole, because the information contained is for the same period. Analyses of the anthracite and Leyden lignite used are given in Tables 4 and 9. Table 9 shows the analyses of the limestone and lime. Tables 3 and 9 give the ore analyses.

All power was measured on the primary circuit and the figures given include furnace, motor, lighting, and laboratory loads, which are designated here as "plant". The power consumption outside of the furnace load is conservatively estimated a t 10,000 kw-hr. per mo. I n Table 10, a deduction of 10,000 kw.-hr. per mo. has been made in figuring' the power consumption of the furnace only, designated as "furnace."

The operating results of October, November, and December are representative of what can be expected n~etallurgically. A higher time operating efficiency, load factor, and lower electrode consumption can be obtained, as was demonstrated in January, but this month is not included because anthracite was the reducing agent.

The gradual decrease in percentage of manganese in the ore charged t o the furnace was partly due to an actual drop in the high-grade ore and partly to the use of Leadville ore. There was no apparent difference in power 'consumption when smelting the lower grade ore, or when producing 70 per cent. ferromanganese instead of the 80 per cent. grade. This was probably due to varying furnace conditions.

The effect of changing the reducing agent from anthracite to lignite was the increased recovery resulting from the change. Thease of an- thracite in January, after the furnace operation was thoroughly estab- lished, resulted in a recovery of 65.4 per cent. of the manganese, compared with 75.5 per cent. recovery in December with lignite. The average manganese recovery for the three months operation with anthra- cite was 60.1 per cent., the average recovery for six months of operation with lignite was 71.1 per cent. The conclusion may be drawn that the use of lignite in the electric furnace production of ferromanganese results in a 10 per cent. higher recovery than when anthracite is the reducing agent.

TABLE 5.-Metallurgical Data: Furnace No. 1, May 1 to Dec. 26, 1918;Furnace No. 2, Dec. 26, 1918, to Jan. 31, 1919

1 May 1 June / July I Aug. 1 Sept. / Oct. 1 Nov. ( Dec. Ian. -- .. - I .. --

Charge: I Manganese ore. high grade, pounds.. . . . . . . . . . . Manganese ore, Leadville, pounds. ............ Coal, anthracite, pounds.. . . . . . . . . . . . . . . . . . . . Coal, lignite, pounds.

I . . . . . . . . . . . . . . . . . . . . . . . .

Limestone, pounds. . . . . . . . . . . . . . . . . . . . . . . . . . Lime, burnt, pounds. . . . . . . . . . . . . . . . . . . . . . . . . Fluorspar, pounds.. . . . . . . . . . . . . . . . . . . . . . . . . . Steel turnings, pounds. . . . . . . . . . . . . . . . . . . . . . .

Average analysis of ore, dry. Mn, per cent. . . . . . . I Average analysis of ferromanganese:

Manganese, per cent. . . . . . . . . . . . . . . . . . . . . . . . . I . . . . . . . . . . . . . . . . . . . . . . . Silicon, per cent.. . . . . . . . . . . . . . . . . . . . . . . . Sulfur, per cent.. : : : : /

. . . . . . . . . . . . . . . . . . . . . . Phosphorus, per cent.. /

Average analysis of slag: I Manganese, per cent. . . . . . . . . . . . . . . . . . . . . . . . . 12.1 12.85 9.14 11.0 9.78 11.36 12.98 12.2 14.44 8i01, per cent. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . : 27.8 . 28.1 1 29.0 ( 31.4 29.8 1 27.7 28.7 29.3 29.7 CaO, per cent.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39.2 39.0 40.6 38.5 41.5 39.2 36.9 37.2 35.2

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mg0,percent 1 5.3 3.2 ; 3 .1 2.9 3.5 2.3 , 2.0 2.8 2.1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . AlrOl, per cent . . 9.5 1 9.7 12.5 10.0 1 8.6 7.5 8.2 7.9 8.5

Ferr~man~anese, pounds.. . . . . . . . . . . . . . . . . . . . . . 144,684 182,970 : 174,700 191.609 193,827 311,180 298.125 278.007 308.167 Ferromanganese, gross tone. . . . . . . . . . . . . . . . . . . . . ! 64.6 : 81.7 78.0 1 85.5 86.5 138.9 133.3 124.2 337.8 blag,pounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1234,337 294.189 1217,845 247.180 225.612 365,042 385.640 372,059 531.875

. . . . . . . . . . . . . . . . . . . . . . . . . . Ratio ore to metal.. ( 3.55 3.45 3.05 , 2.77 2.95 2.85 3.05 2.85 3.16 S l ~ p e r ~ o s s b n m e t ~ l , p o u n d a . . . . . . . . . . . . . . 3,627 3,6001 2,792 2 , 8 9 2 / 2.608 2.626 2.890 2.995 3.859 Slag per net ton metal, pounds. . . . . . . . . . . . . . . . . .I 3.241 1 3,215 2,492 2,580 2.328 2.347 2.588 2.676 3.446

. . . . . . . . . . Total manganese in metal, per cent. :. . I 57.5 57.5 1 8 5 . 8 . 75.4 i 70.2 73.45 67.0 75.5 65.4 . . . . . . . . . . . . . . Total manganese in slab, per cent.. 14.1 15.0 9.5 13.6 1 1 0 . 0 13.15 15.85 16.28 22.5

. . Total mangane~~e unaccounted, per c e n t . . . . . 28.4 / 27.5 1 24.7 11.0 19.2 13.55 17.15 8.22 12.1 I

NOTB.-All weights are weights a s weighed. Phase voltage 70-75 volts Average load on furnace about 1100 kv-a.

TABLE 6.-Consumption of Furnace Supplies

' >lay ' June July I Aug. Sept. Oct. Sov. Dec. Jan. .-

E l e ~ t r o d e s , ~ o u n d s . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.477 24,504 22.257 23,596 20.139 26,779 24,811 28.457 1!4,41.i Elertrodes per gross ton metal, pounds. . . . . . . . . 409.8 300 , 285.3 277 232 103 186.5 229 141 Elertrode uaed. diamrter inches.. . . . . . . . . . . . . . . . 16 8. 16 1 1 6 1 7 14-17 14-17 14-17 14-17-24 24 Electrode paste, pounda.. . . . . . . . . . . . . . . . . . . . . . . 300 200 b b b 1 b b b

Electrode paste per gross ton metal, pounds.. . . . . , 4 . 6 2.4 Tappin ,~s tee l ,pouuds . . . . . . . . . . . . . . . . . . . . . . . . 2.124 3,347 2.318 1,074 1,349 1.722 4.786 2.011 1.273 Tappiug steel per cross ton metal, pounds.. . . . . . . 33.1 40.9 29.8 12.5 15.6 12.3 36.3 16.3 9 . 3 Oxygen, cubic fee t . . . . . . . . . . . . . . . . . . . . . . . . . . . . None 2.000 2,000 2.000 800 Sone S o n e 400 Sone Oxygen per moss ton metal, cubic feet. . . . . . . . . . . . 2 6 . g i 25 23.4 9 . 2 ! 3.2 JIagnenite. pounds.. . . . . . . . . . . . . . . . . . . . . . . . . . . Sone None None 675 3.152 None Xone None Sone JIagneaite per moss ton metal, pounda.. . . . . . . . . . 8 , 36.5 I Fireclay, pounds.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.531 ' 4.687 2.202 8,000 8.370 ! 11,818 13.275 8.765 9.353 Fircclay per gross ton metal, pouuds. . . . . . . . . . . . . 85 .6 57.3 28 93.5 97 ! 85 99.5 7 0 . 5 , 68 Pitch, pounds.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sone None None 275 1,225 Sone None Sone Sone

. . . . . . . . . . . . . Pitrh per cross ton metal, pounds.. 3.2 14 I !

s

16-in. clertrode not available, used 4-8-in. electrodes per holder. Included in weight of electrodes.

TABLE 7.-Power Consu?nption

. . . . . . . . . . . . >laximum I-hr. demand, kilowatts.. . . . . . . . . . . Total kilowatt-hours consumed.d plant

. . . . . . . . . . . . . . . . . . Plant load factor, per cent . . . . . . . . . . . . . . . . . . . . Percentage of time operated.. ,

. . . . . . . . . . . . . . . . . . . . . . Furnace days operated.. . . . . . . . . . . . . . . J le ta l per calender day, pounds..

. . . . . . . . . . . . . . Metal per furnace day, poundn.. . . . . . Kilowatt-hours per cross ton meta1.d p lant . . . . . . . Kilowatt-hours per short tond metal, plant. .

J l ay e June' July a Oct. Sov. Jan.'

(d) Shut down 7 days, no electrodes. (d) Includes miscellaneous plant power, approximately 10.000 kw.-hr. per mo. ( b ) Down 3 days, no electrodes. All power measurements made on primary aide of transformers. LC) Using anthracite coal. NOTE.-Average load on furnace about 1100 kw. Phase voltage 70 t o 75 volte.

28 MANUFACTURE OF FERROMA~GANESE IN THE ELECTRIC FURNACE

The use of lime instead of limestone did not result in a lower power consumption as it should do theoretically; this was probably due to the limestone being calcined on the top of the furnace by heat which other-

I

wise would be wasted. The furnace voltage during the last four months of operation was 70

to 75 volts and in May and June, i t was 75 to 80 volts; there appeared t o be a considerably higher volatilization loss a t this time. When the voltage was increased to 120 volts for several days, the volatilization was marked, and the furnace load was difficult to control. A voltage of 65 t o 75 volts seems to be satisfactory for a 1100-kw. furnace.

The recovery of manganese during the last quarter of 1918, in which the smelting practice was thoroughly established and i t was possible t o operate for a full month, averaged 72 per cent. Of the loss, 28 per cent., 15 per cent. was in the slag and 13 per cent. unaccounted loss. The average slag volume was 2830 Ib. (1283.7 kg.) of slag per gross ton of metal. The percentage manganese in the slag averaged 12.2 per cent. As this proved to be about as low an average slag as could be produced without excessive formation of carbide and a decrease in output, the recovery on this ore cannot be increased except by reduction of the un- accounted loss, which might be accomplished by lowering the voltage or briquetting the fine ore. The average recovery over the nine months period of operation, including good and bad months, was 68.6 per'cent. Of the loss of 31.4 per cent., 13.4 per cent. was in the slag and 18.0 per cent. unaccounted loss. The average slag fall for the nine months was 3121 lb. of slag per gross ton of ferromanganese.

The average "plant" power consumption for the last quarter of 1918 was 5066 kw.-hr. per gross ton of ferromanganese, or 4524 kw.-hr. per short ton. The average "furnace" power consumption was 4990 kw.-hr. per gross ton or 4457 kw.-hr. per short ton of ferromanganese. This power consumption can be met regularly with the grade of ore smelted: 34.94 per cent. manganese and 13.2 per cent.' SiOz, in an 1100 kw-hr. furnace. I t is possible that a lower power consumption would have risulted in January except for the use of anthracite instead of lig- nite. The average "plant" power consumption for the whole nine months of operation was 6240 kw.-hr. per gross ton of ferromanganese or 5586 kw.-hr. per short ton.

For the four months after September, until it was closed down, the 1 1 "plant" operated at an electrical load factor of 81.6 per cent.; the fur-

nace" load factor would be a little higher than this. The average per- centage of time operated was 89.1 per cent. The power factor, as determined by the power company, varied between 85 and 90 per cent. This load factor was attained with hand operation of electrodes. Some typical load curves are shown in Figs. 13 and 14. If operations had been continued the load factor would have averaged 85 per cent., and

. .

ROBERT .M. KEENET AND JAY LONERGAN 29

FIG. 13 . -24-~~. LOAD CURVE OF FIG. 1 4 - 2 4 ~ ~ . LOAD CURVE OF FURNACE No. 1. K = 1.92. FURNACE NO. 2. K = 1.92.

the percentage of time operated 90 to 95 per cent. The jump from a load factor of 79.9 per cent. in December to 85.1 per cent. in January, and the jump from 85.3 to 96.4 per cent. of time operated was due to the use of 24-in. electrodes instead of the smaller sizes, and better mechanical ar- rangements for changing electrodes on furnace No. 2.

The electrode consumption for the last four months of operation averaged 187 Ib. (84.8 kg.) per gross ton of ferromanganese, or 162 lb. per short ton, or 32 lb. (14.5 kg.) per 1000 kw.-hr. passing through the electrodes. During this time 14-in., 17-in. and 24-in. electrodes were used. Except when using 24-in. electrodes, there was always one 14-in. electrode in thefurnace but never more than one. About one half of the 17-in. electrodes were home-made "butt" electrodes produced a t the plant from crushed broken electrode. ~ h e s e electrodes had 6 to 8 hr. greater life than the best purchased electrodes. This was because the

3

+ 4 .. . .

6 .- '.

1 Z 3 4 5 6 7 8 Silicon, Percent

9

FIG. 18.-EFFECT OF VARIATION OF SILICON ON CARBON IN FERROMANGANESE.

material from which they were made was better material for electrode manufacture than that used in the purchased electrodes. I t was more highly graphitized because it had been electrically calcined twice.

The beneficial effect of operating a furnace with the electrodes carry- ing their normal load or less is shown by the results of the January opera- tion of furnace No. 2 with 24-in. electrodes. The electrode consumption for January averaged 141 lb. (63.9 kg.) per gross ton of ferromanganese, or 125 lb. per short ton, or 26 lb. (11.8 kg.) per 1000 kw.-hr. These elec- trodes were oversize for the load carried. A threaded electrode should not be operated with a load of over 30 amp. per sq. in. of cross-section, although a "butt" electrode will carry 50 amp. per sq. in. If threaded electrodes are overloaded, there is not only an excessive electrode consump- tion, due to the joint corroding and finally breaking, but much operating time is lost. With the 24-in. electrodes, no part of the electrode was lost except the connecting pin, which would drop out, but with the 17-in threaded electrode it was common to lose 2 ft. of the electrode. It was as economical to use 17-in. butt electrodes as threaded 17-in. electrodes.

ROBERT M. KEENEY AND JAY LONERGAN 31

The average load carried by these electrodes was 10,500 amp. On this basis the electrodes were loaded as shown in Table 8.

TABLE 8.-Current Carried by Electrodes - - -

Diameter of Electrode. Inches

8 (four per holder) 14 16 17 24

- -

Cross-section 1 Area. Square Inches Approximate Amperes per

Load, Amperes Square Inch

TABLE 9.-Analysis of Materials Charged and Tapped i n Last Quarter of 1918

1 Manganese Ore, Dry, Per Cent.

TABLE 10.-Summary of Operating Results i n Last Quarter of 1918

-

. . . . . . . . . . . . . Manganese. Si01 . . . . . . . . . . . . . . . . . . . . MgO . . . . . . . . . . . . . . . . . . Iron . . . . . . . . . . . . . . . . . . . . CaO.. . . . . . . . . . . . . . . . . . . AlsOs . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . Phosphorus.. Sulfur.. . . . . . . . . . . . . . . . . . BaO.. . . . . . . . . . . . . . . . . . . Moisture. . . . . . . . . . . . . . . . Ash.. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . Volatile matter.. Fixed carbon . . . . . . . . . . . CO, . . . . . . . . . . . . . . . . . . . COI and HtO:. . . . . . . . . . . Copper. . . . . . . . . . . . . . . . . . Carbon. . . . . . . . . . . . . . . . . . Silicon.. . . . . . . . . . . . . . . . .

Manganese ore, pound. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Coal, lignite, pound.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Limestone, pound.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Lime, burnt, pound.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Steel turnings, pound.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ferromanganese, pound.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ferromanganese, gross tons.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Slag,pound . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ratio ore t o metal..

Slag per gross ton metal, pound.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Slag per net ton metal, pound.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Total manganese in metal, per cent.. . . . . . . . . . . . . . . . . . . . . . . . . . . Total manganese in slag, per cent. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

=bite COul. z;t.

- -

34.8 13.2 0.4 7.1 5.7 3.0 0.038 0.61 2.3

7.09

Lime-

'"p",","' Cent,

Lime, Per

Cent.

Steel Turn- Ings, Per

Cent.

Ferro- manganese, Per Cent.

Slag, Per

Cent.

TABLE 10.-Summary of Operating Results in Last Quarter of 1918 (Continued)

Total manganese unaccounted, per cent.. . . . . . . . . . . . . . . . . . . . . . . Electrodes,pound . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electrodes per gross ton metal, pound . . . . . . . . . . . . . . . . . . . . . . . . Electrodes per net ton metal, pound.. . . . . . . . . . . . . . . . . . . . . . . . . . Electrode per 1000 kw-hr., pound.. . . . . . . . . . . . . . . . . . . . . . . . . . . . Tapping steel, pound.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Tapping steel per gross ton metal, pound.. . . . . . . . . . . . . . . . . . . . . . Oxygen,cubicfeet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Oxygen per gross ton metal, cubic feet.. . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fireclay, pound.. Fireclay per gross ton metal, pound.. . . . . . . . . . . . . . . . . . . . . . . . . . Average maximum demand of plant". . . . . . . . . . . . . . . . . . . . . . . . . . Total kilowatt-hours consumed in planto.. . . . . . . . . . . . . . . . . . . . . . Average of plant, load factor per cent.. . . . . . . . . . . . . . . . . . . . . . . . . Average time operated, per cent. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Furnacedaysoperated . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Metal per calendar day, pound. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Metal per furnace day, pound. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Kilowatt-hours per gross ton metal, plant. . . . . . . . . . . . . . . . . . . . . . Kilowatt-houn per short ton metal, plant.. . . . . . . . . . . . . . . . . . . . . Kilowatt-hours per gross ton metal, furnace only.. . . . . . . . . . . . . . . Kilowatt-hours per short ton meta!, furnace only.. . . . . . . . . . . . . . .

Includes approximately 30,000 kw .-hr. consumed in general plant operation.

Impurities in Ferromanganese

Elimination of impurities is not one of the problems of ferroman- ganese manufacture. As 0.2 per cent. phosphorus is allowable in ferro- manganese, it is possible to use almost any ore by mixing it with a low- phosphorus ore. Sulfur is readily eliminated by the strong reducing atmosphere of the furnace, the basic slag, and the action of manganese carrying sulfur into the slag as manganese sulfide. Most of the sulfur that appears in analyses is probably from particles of slag in the sample. No attention is paid to the carbon content of the metal, because a low carbon alloy is seldom required. It fluctuates as shown in Table 11, decreasing with increased silicon in the metal.

Several electric-furnace plants make it a practice to produce ferro- manganese containing 5 per cent. silicon, claiming that this is economical because it reduces the manganese loss by reduction of slag volume; the authors, though, cannot see the economy of producing metal containing over 2 to 3 per cent. silicon.

A series of tests made during plant operation showed that 1.15 per cent of the sulfur charged remained in the ferromanganese. Of the 98.85 per cent. eliminated, 39.35 per cent. passed into the 'slag, and 59.5 per cent. was lost in dust or volatilized. Of the phosphorus charged 52.6

ROBERT M. KEENEY AND J A Y LONERGAN 33

per pent. remained in the metal. Of the 47.4 per cent. eliminated, 20.81 per cent. was in the slag and 26.59 per cent. was lost in dust or volatilized.

TABLE 11.-Effect of Increase of silicon on Percentage Carbon in Ferromanganese

I t is possible to give the complete cost data of the plant described as the operating corporation has gone out of business. 'fhese'costs are wartime costs. Table 12 shows the unit costs for the last four months of operation, and Table 13 the cost ,per gross ton of ferromanganese produced during this period. The labor employed was as follows:

Silicon. Carbon. per cent . 1 per cent .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Master mechanic.. : Blacksmith . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Unloading foreman. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Unloading labor..

. Truckdriver . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Night foreman.

Furnace foremen. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Head charger and tapper..

Furnacelabor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Total . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Manganese, Per Cent.

0.12 5.79 ! 8 0 . 6 0.30 5.89 1 77.62 0.34

5.45 1 ;::y 0.37 5.90 0.52 . 6.01 83.89 0.98 ' 6.17 1 79.21 1.10 6 . 0 6 1 80.55 1.24 5.52 79.84 1.48 5.12 1 82.65 1.50 5.78 1 81 :4 1.57 5.35 79.15 1.97 , 4.97 , 79.4 2.10 5.07 , 79.65 2.48 5.35 i 81.77 2.70 ' 5.07 79.85

. 2.72 4.25 79.46 2.95 .4.30 79.79 3.05 5.13 1 79.52

Silicon, Carbon. I Per Cent. Per Cent. Manganese. Per Cent.

--

I 3.47 j 4.15 78.57

78.56 74.7 78.84 79.92 80.19

, 77.75 77.16 78.97 78.28 78.16 79.15 75.18 76.53 75.31

3.55 4.37 5.08

4.09 4.69 4.11 4.42

5.02 :::: 4.69 4.84. 5.18 5.17 4.04

8.07 ' 3.28 73.44 8.16 1 3.60 1 75.64 8.22' i 3.49 j 75.20

5.56 5.63 - 5.68 6.06

4.53 4.54 4.30 3.76

6'.32 1 4.31

34 MANUFACTURE OF FERROMANGANESE I N THE ELECTRIC FURNACE

TABLE 12.-Unit Costs

Cents Cents Cents Cents

Manganese ore, high-grade, per pound. ; Manganese ore, Leadville, per pound. . I

. . . . . . . . . . . . Lignite coal, per pound.. . . . . . . . . . . . . . Burnt lime, per pound.. . . . . . . . . . . . . . . Limestone, per pound..

. . . . . . . Carbon electrodes, per pound.. / Power, per kw.-hr. . . . . . . . . . . . . . . . . . . I

Tapping steel, per pound.. . . . . . . . . . . . . . . . . . . . . . . . . . . . Fireclay, per pound..

Steel turnings, per pound.. . . . . . . . . . . . . . . . . . . . . . . Anthracite coal, per pound..

TABLE 13.-Cost Per Gross Ton of Ferromanganese

The manufacture of ferromanganese in the electric furnace was the result of war demand for ferromanganese and the need of treatment of domestic ore in order to make available for war use a considerable tonnage of shipping used for transporting ore from South America. The industry

1 October. , 1918 - -

November, ' ~ecember . January. Average 1918 1 . 1918 I 1918 1 per Month

Manganese ore.. . . . . . . . . . . . Coal . . . . . . . . . . . . . . . . . . . . . Limestone and lime. ....... 2.47 4.58 Carbon electrodes.. . Tapping steel.. ........... 0.97 !

Fireclay.. ................ , 0.40 1 0.50 . 0.32 ! 0.30 . . . . . . . . . . . . Steel turnings. 0.02 !

Miscellaneous supplies.. . . . 0.08 ' 0.76 : 1.21 ! Power.. . . . . . . . . . . . . . . . . .# 22.65 . 23.35 1 25.30 ) 23.44 Labor. . . . . . . . . . . . . . . . . . . . ' 20.12 1 21.66 23.24 / 21.51 Supervision.. . . . . . . . . . . . . . . 4.52 . 8.14 1 10.62 9.20 Repairs and renewals, sup-

plies.. ................. ' 1.77 2.32 : 1.83 ; 1.07

0.51 23.68 21.63

8.12

1.75 Repairs and renewals, labor 1.92 1.99 1.84 2.11 1.96 Truck supplies.. .......... 0.37 ; 0.31 0.48 , 0.52 / 0.42

. Truck, labor.. ............ 0.67 1 0.70 0.92 0.87 ...... Laboratory, supplies.. 1.56 0.46 , 0.53

Laboratory, labor.. . . . . . . . 1.95 i 3.16 3.30 2.86 Administration.. . . . . . . . . . 6.64 i 9.96 11.19 / 4.24 8.00

! I

i $168.68 1 $182.19 $190.52 $164.28 $176.41 i

ROBERT M. KEENEY AND JAY LONERGAN ' 35'

. . reached.its peak in 1918, with a yearly production of 23,000 tons valued .

at about $5,000,000, or 7 per cent: of the totalferromanganese production of the United states. During 1919, there, mas little'electric furnace , ' .

production, but operations, were resumed early in 1920, with an esti- mated production for the year of 10,000 tons. . Nine plants in this country' have produced ferromanganese in the' electric furnace, all but one. of which were constructed for that purpose. There are installed 33 electric -furnaces of from 350 kv-a. to 5000 k,v-a. capacity, with a total installed .

transformer capacity of 58,000 kv-a. The future of the process, rests in '

its,use as a method of ore treatment by the owner of a manganese mine with other favorable local conditions. . .

Comparing blast-furnace production of ferromanganese with electric- , .

furnace production, the paper shows that the same recov'ery of manga- ' .

nese, 72 per cent., was made in the electric furnace smeltingof orecon: taining 34.8 per cent. manganese and 13.2 per cent.Sio2, as in the blast' ,

furnace smelting of ore containing 40.33 per cent. manganese and 8.6 per cent. SiO2. The slag and stack l&ses were practically the same in

. each case. The electric furnace consumes approximately one quarter of the quantity of carbon required by the blast furnace. It is probable that on the same.grade of ore, the electric furnace would show at least 5 per cent. greater.recovery than the blast furnace. . .

' In smelting oxide ore, better operating results were obtained when ' . . any necessary iron was added in the form of manganiferous ore rather'

than as steel turnings. There was no economy in the use of lime as a flux instead of limestone. Lignite coal proved to be a better reducing agent than anthracite, bituminous coal, or coke, alt,hough coke was not given a long trial. From the viewpoint of output, power ~onsumpt~ion, and labor costs, intermittent charging gave better results than con-, .

,tinuous charging. Good results were obtained by tapping into cast-iron. pots instead of beds. . . .

Some of the operating difficulties t'hat may be encountered are: ' -

formation of carbide in the furnace, -blowing, failure of bottom, tapping, \ '

failufe of electrode holders, and bridgihg of charge. Of thesetroubles , '

formation of carbide is the most serious. The tendency to form carbide . can be decreased by the use of lignite coal as a reducing agent;.by.not

attempting to produce a slag containing much less than 12 per cent. . manganese; and by the use of a low-carbon flushing charge'when carbide

' '

begins to form. Blowing is decreasedby intermittent charging, and by - 'elimination of fine material in the charge. The authors prefer the mag- nesite bottom to the carbon bottom. Failure of electrode holders is . . a matter of design. ~ r i d ~ i n ~ of the charge. is eliminated by proper spacing of electrodes.'

Manufacture of ferromanganese containing 73.6 per cent. manganese ..: was accomplished i n an 1100-kw. electric furnace by smelting ore con-

taining 34.8 per cent. manganese and 13.2 per cent. SiOs with an average power consumption of 4990 kw.-hr. per gross ton, or 4457 kw.-hr. per short ton. There was tapped 2830 lb. (1283.7 kg.) of slag per gross ton of ferromanganese. The electrode' consumption averaged 202 lb. per gross ton of metal, but during the last month of operation averaged 141 lb. per gross ton, or 125 Ib. (56.7 kg.) per short ton, or 26 lb. per 1000 kw.-hr., which is what can be attained regularly. The plant operated 89.1 per cent. of the total time during the last four months with an electrical load factor'of 81.6 per cent. using hand regulation of elec- trodes. The load factor for the last month was 85.1 per cent., and ;the percentage of total time operated was 96.4 per cent., figures which can be made regularly by experienced operators.

The average cost of production, f.0.b. plant, exclusive of taxes, interest, depreciation, and sales expense, during the period of four months when materials on hand permitted continuous operation of one 1100- kw. furnace was $176.41 per gross ton. Excluding reduction of costs due to cheaper ore supply, cheaper power, better design of plant, and higher operating efficiency, considerable reductions in cost of production of ferromanganese in the electric furnace are possible if the Soderburg continuous feeding electrode now installed in one plant continues to be a success, and if the unaccounted loss of manganese can be decreased.