composition. The latter evil is especially aggravating in malleable work where the importance of keeping the two elements sulphur and manganese, in a proper relationship is well recognized.

Used as a Mixer

In order to overcome these last two disadvantages we might utilize some sort of a mixer that would act as a reservoir and by holding a given quantity of metal allow it to become

122

FIG. 1-RESULTS OBTAINED WITH TEST WEDGES OF LOW MANGANESE IRON

constant in temperature and composition. The difficulty with this plan, is of course, that the temperature of the molten metal would be constantly lowering unless some external source of heat were applied, so that the question narrows itself down to just what this source of heat would be. The electric furnace, at once suggests itself as ideal, for in addition to supplying the heat needed, it will also allow the removal of the greater part of the sulphur by use of a proper slag. Not only this, but it will, in addition, allow an iron of any carbon con

tent to be made by the means of additions of either cold or liquid steel, so that compositions of any range of carbon and silicon can be made and any castings of sections practicable in malleable iron may be poured.

The advantages of such a duplex process are readily apparent to anyone familiar with the production of cupola malleable iron. Cupola iron is especially suitable for castings of light section, such as pipe fittings, etc. In the latter case where an iron high in carbon is desirable, because of the greater ease of threading the fittings this process should recommend itself immediately. In addition to the advantages previously enumerated it would permit the annealing of the white iron at a considerably lower temperature than is practicable with a high sulphur iron, such as is ordinarily produced in the cupola.

Results of Experiment

In order to determine how far it is possible to reduce the sulphur in cupola iron in a reasonable length of time by this process, some sprues and scrap were obtained from a manufacturer of cupola malleable iron and melted in a small Heroult electric furnace. The material had the following composition: Silicon, 0.65 per cent; manganese, 0.53 per cent; phosphorus, 0.143 per cent; sulphur, 0.20 per cent, and carbon, 3.06 per cent. This analysis was supplied by the firm which furnished the scrap and was given as the average. No attempt was made to check it due to the difficulty of average sample of material of this nature.

securing an

The scrap was melted bare in the furnace and then a basic slag of lime thinned with spar was put on. Finely ground petroleum coke was sprinkled over the slag to reduce all oxides throwing the metals back into the bath, the slag turning a creamy white color, indicating the completeness of these reactions. Calcium carbide is formed in the slag as is evidenced by the pronounced odor of acetylene when such a slag is moistened in water. A slag of this character will readily absorb sulphur from the bath and while the reactions require some time for completion, the practical elimination of the sulphur is very rapid.

Accordingly, 15 minutes after the slag was made a sample was taken for analysis, with the following results: Silicon, 0.57 per cent; manganese, 0.54 per cent; sulphur, 0.057 per cent; and carbon, 3.36 per cent.

Note the reduction of sulphur which is very striking when it is considered that the sample was taken just 15 minutes after the formation of the slag. There is something further of note,

namely the retention of the manganese. This element is readily oxidized in the cupola and it is customary to run the mixtures very high in manganese to take care of what is burnt out in melting and still leave enough for the high sulphur that is obtained. With the electric furnace, no speigeleisen, ferromanganese, or high manganese pig is necessary, and in fact, as will be later shown, the mixtures, to get best results, would have to be kept low in this element. As was mentioned before, the analysis before melting was supposed to represent the average

so that the slight difference of silicon obtained need not be considered. The carbon, however, seems to have increased, but there is a possibility of doubt here, as 3.06 per cent seems rather low for cupola iron in the first place. Then again, when carrying a refining slag on low carbon steel, which is greedy for carbon, the increase due to additions of coke to the slag are so small as to be negligible.

At the same time the sample was poured for analysis, a set of test bars was poured and after cooling the bars were broken and fractures examined. The bar 5%-inch in diameter was clear white and the 14-inch bar nicely mottled so that the composition seemed all right for the class of material under consideration. A set of test bars and wedges for annealing was then poured.

Steel and Ferrosilicon Added

In order to obtain iron more suitable for work of a heavier section, some steel and ferrosilicon in calculated amounts were added to the bath. The calculation was based on the 3.08 per cent of carbon as supposed to have been in the original material but as it was actually higher in carbon the result was higher than desired. However, it was near enough for the purpose, to illustrate the possibility of producing an iron of a lower carbon than is ordinarily obtained in the cupola. The actual analysis is given as follows: Silicon, 0.75 per cent; manganese, 0.53 per cent; sulphur, 0.036 per cent; and carbon, 2.90 per cent.

It is interesting to note in passing that the analysis obtained was exactly that expected, taking into consideration the actual first analysis, and the analysis and amounts of the additions. Anyone who has had experience in adding ferrosilicon to an air furnace where a strongly oxidizing condition prevails, to bring up the silicon, can appreciate this. The sulphur is still lower in this sample, which can be attributed to the further refining of the slag, for even had the steel contained no sulphur the resultant would have only been 0.045 per cent, had the slag not absorbed more from the bath.

Test bars of this composition were poured and both sets of bars were subsequently sent to the firm that supplied the

scrap for annealing. They were packed in the second pot from the bottom and placed in about the middle of the oven.

The examination of the bars after annealing showed that the metal was inclined to be rather "short" and that it lacked toughness. This might easily be expected when the high manganese, low sulphur composition is taken into consideration. The fractures viewed with the eye looked like typical blackheart iron with the exception that the rim of ferrite is heavier

than usual. Under the microscope, however, we find that the structure consists of a rim of ferrite and the interior a matrix of ferrite and pearlite embedded with temper carbon. instead of a pure ferrite matrix, characteristic of good blackheart iron.

Sulphur Eliminated

As it was impossible to obtain any low manganese, high sulphur cupola iron without making it especially for our