ing conditions with the slag resulting from the melting-down operation. The acid process is, therefore, a one-slag process, but it is necessary to purchase scrap with the phosphorus and sulphur content below the maximum allowed in the finished steel, and in most cases the economy effected by a oneslag operation, compared with the basic process, would be offset by the premium paid for low-phosphorus scrap. To produce steel of a uniformly high quality by the electric process we have only to consider the production of a uniformly deoxidized silicious slag.

Unlike the basic process from which we can produce slag of a definite composition and a definite volume, the acid slag varies considerably, both in analysis and quantity. The controlling feature in the operation of the acid process seems to be the quantity of iron-oxide present at the time the heat is melted, irrespective of whether this oxide is introduced with the steel scrap as rust, or whether it results from oxidation due to air leaks in the furnace, or from the addition of iron ore. The slag volume in the acid process is controlled directly by the action of this iron oxide and silica.

Unfortunately, scrap in any scrap yard is variable with regard to rust. A small coating of rust on light scrap such as turnings, which is, in certain proportions, a desirable scrap for an electric furnace, represents a much greater percentage of the scrap than the same coating of rust on heavier scrap.

Oxidation Due to Leaks

The amount of oxidation due to furnace leaks is a variable with every heat, so that oxidation from this source is not necessarily at all consistent. Scrap charged in the evening and allowed to stand in the furnace until morning, or charged Saturday night and allowed to stand until Sunday or Monday will become oxidized, the amount of oxidation. depending upon the weight of the individual pieces of scrap and the care which was used in sealing the furnace after it was charged. It, therefore, follows that the amount of iron oxide in a furnace during melting is a variable which it is impossible to estimate. With the iron oxide varying between.

heats, the slag volume will be similarly affected and there is no means of determining, before the heat is tapped, what percentage of slag volume there is.

The reducing conditions necessary for the production of the highest grade steel and for making heats within the specifications desired require a slag with a definite composition. To secure the proper slag without delaying operations, its composition must be judged solely by its appearance, and the fact that this slag is not obtained through reducing conditions alone makes the acid electric process one which requires more skill and judgment on the part of the furnace man than the basic process.

The acid slag after melting and before any additions are made is principally a complex silicate of iron and manganese, containing usually about 50 to 60 per cent silica, the remainder being principally oxides of iron and manganese. Practically all the manganese of the charge with the exception of about 10 per cent passes into the slag as oxide. The percentage in the slag, of course, will depend upon the slag volume.

The oxidation of the bath, and of the subsequent additions of manganese, appears to be controlled by the percentage of FeO in the slag. Whether the excess FeO is combined in some form of silicate, or whether there is free iron oxide I do not know, but whatever form the iron oxide is in, it acts as an oxidizing agent on carbon, silicon and manganese.

Further, this iron oxide cannot be appreciably or satisfactorily reduced from the slag by adding coal or coke dust as in the basic process for two reasons. First, a silicate in which the acid constituent is in excess is more difficult to dissociate than a silicate in which the bases are in excess, as in the basic process, and further the iron is more stable in a silicate in which the iron base is the predominating base, as in the acid process. In the basic process, with CaO and MgO the controlling bases in a basic silicate, iron oxide, whether free or combined with the other bases as a complex silicate, is easily reduced by coal or coke.

The second reason why it is impractical, and perhaps impossible to reduce FeO from the acid slag by coal or coke is that at the high temperatures in the vicinity of the arcs, SiO2 is reduced by carbon. This phenomenon is familiar in the bessemer process, where, when a certain high temperature has been reached, silicon in the bath increases during the blow.

In the acid process, therefore, if the slag is subjected to reducing conditions before it is of the proper composition, silicon will be reduced and the bath will contain a percentage of silicon varying with the intensity of these conditions and their duration.

As the iron oxide is reduced with difficulty, it must therefore be replaced by a stronger base. This is partially accomplished by the addition of lime. Manganese ore, if free from iron, should also be used for the same purpose. It is by replacing iron in the slag with calcium and manganese that it is possible to produce the necessary reducing conditions and secure uniform retention in the bath of subsequent manganese and silicon additions. Fortunately a slag with the proper percentage of lime and manganese can be determined by its appearance.

Similar to an Open-Hearth Slag

An oxidizing acid slag such as obtained after melting down is black and opaque or slightly glassy similar in all respects to a common acid open-hearth slag which it really is, chemically and physically. As soon as a heat is melted and the carbon is observed to be sufficiently low, lime should be added immediately up to 20 or 30 per cent of the slag volume, also it is preferable to add manganese ore sufficient to give 10 to 15 per cent MnO in the slag. When the slag contains. sufficient lime for practical purposes it will no longer be black but will turn to a green or robin's egg blue and when cooled in the air, the surface will turn black or brown depending upon the proportions of iron and manganese oxide, black if it contains too much iron oxide, and brown if the iron oxide. is sufficiently removed. Coke dust or coal should be used dur

ing the interval of adding lime and manganese ore, and until the heat is tapped to prevent any further oxidation of iron. If all the constituents of the slag are actually combined with nothing in suspension, and the furnace is free from air leaks and working under a normal reducing atmosphere, a translucent, almost transparent, glassy slag of yellowish green or greenish blue is obtained. With this transparent slag and the greenish blue slag coated with a chocolate brown, reducing conditions are the best that can be obtained by the acid process. Subsequent additions of manganese and silicon will result in uniform losses and therefore give uniform results in the finished steel with respect to these elements.

A typical analysis of a proper finishing slag is as follows: SiO., 58.70 per cent; CaO, 21.25 per cent; MnO, 12.01 per cent; FeO, 3.10 per cent; MgO, 1.15 per cent; Al2O3, 3.35 per cent; P, 0.003 per cent; and S, 0.005 per cent.

If the slag during these operations shows a tendency to be porous or spongy it is still in an oxidizing condition. A proper acid slag should be free from sponginess or indications of gas pockets.

Three Alternatives

Three alternatives to the ordinary scrap melting, one-slag, process can be used to overcome the variables introduced by rust or other oxidation, namely:

1. Duplexing from a converter, or open-hearth furnace. 2. Slagging-off immediately after the charge is wetted. 3.-Tumbling the scrap before charging to remove rust. The first of these is possible only in few plants. The second, to my mind represents the only satisfactory method of operating the acid process under the conditions found in the majority of electric-furnace plants. The third is expensive, impractical, and does not prevent the variations due to oxidation in melting, but would produce more uniform results than the melting of scrap with variable degrees of rust.

It is important that the carbon content of the bath be as low as, or lower than that desired for tapping, before an effort is made to reduce the iron oxide percentage of the slag.

Iron ore added after the lime and manganese addition, delays the operations and produces a slag which is thin and scorifying.

The yield of steel from charge to ladle is higher in the one-slag acid process, than in the two-slag basic process on account of the losses in the latter due to "slagging off." However, if the acid process were carried out with two slags the losses would be about equal.

Open Questions

Statements that the acid process produces steel with fewer slag inclusions, and of greater fluidity for castings, than the basic, are open questions. My observations of inclusions have led me firmly to believe that foreign inclusions. are of far more frequent occurrence than any possible inclusions resulting from reactions within the steel made by any modern process. If inclusions are found where large masses of steel are not involved, such elements as ladle linings, nozzles, washes, pouring, and molding should be thoroughly investigated before any connection can be laid upon whether a process is acid or basic. From the point of view of inclusions resulting from a process, the basic electric process gives conditions more favorable for their minimum than the acid process.

The fluidity of acid electric steel might be greater on account of the almost invariably higher phosphorus. Unfortunately this feature of the acid process has never been determined. A simple set of experiments carefully carried out could determine this, and it is to be hoped that in the near future this question which is of importance particularly to foundries casting small and intricate shapes, will be answered by actual figures.

No difference has been found in the physical properties of well made acid electric-steel castings or forgings compared to basic electric-steel, and for the purpose of castings, therefore, we may consider the product the same in this respect.

Briefly, we must consider the acid electric-steel process as a slag-making process. Oxidizing conditions caused by rust and by furnace operations influence the quantity of slag, and this quantity is always a variable. Iron oxide in the slag