THE BOWER-BARFF RUSTLESS IRON PROCESS.

BY

GEORGE W. MAYNARD, OF NEW YORK.

THE substitution of costly and not easily oxidized metals for iron in structural, engineering and art work has been rendered necessary because of the difficulty which has always been experienced in so protecting the surfaces of cast or wrought iron as to prevent the corrosive action of the atmosphere or water. The multitudinous devices in the form of paints, coating by immersion in molten metals (galvanizing and tinning), or the electrolytic deposition of metals have at the best been very unsatisfactory expedients.

With the exception of the noble metals, iron far exceeds all others in the ease with which it can be manipulated and converted into artistic forms, as is shown in the beautiful hammer work of Germany and Belgium, and the fine castings of Ilsenburg and Coalbrookdale in Europe, and Poulson and Eger, the Magee Furnace Co., the Yale Lock Manufacturing Co., and others in this country. But rust has come in to spoil many a beautiful work of art, notwithstanding the veneer of copper, brass and nickel, so that it began to look as though iron was doomed except for such constructive work as would be hidden by brick and mortar. It is well known that the architect shuns the use of iron where it comes in contact with stone, owing to the unsightly streaks of rust which speedily deface the surfaces of both the iron and stone. Any device which will absolutely prevent rust, without marring the beauty of the iron must be hailed with delight by the architect, manufacturer and engineer. The coating must admit of easy application, must be pleasing to the eye, must be susceptible of taking and holding paint-where, for any reason paint becomes necessarymust not injure the iron, and must be cheaply and quickly applied. These requirements are complied with in a marked degree by the Bower-Barff Process.

The process has been so long before the public and has been so fully described in the able papers of Mr. George Bower, read before the Iron and Steel Institute of Great Britain in 1881, and before the Society of Engineers in London last month, and by his son, Mr. A. S. Bower, before the American Institute of Mining

Engineers in February last in Boston, that much which I have to say will of necessity be extracted from these papers because they cannot well be improved upon. I here make a general acknowledgment of my obligation to these gentlemen.

In 1876 Prof. Barff read a paper describing his process before the Society of Arts in London, which created a great deal of interest at that time.

The Barff process consists in placing the objects of iron or steel into an iron muffle, which is heated by the external application of heat, and then, when at a sufficiently elevated temperature, depending upon the nature of the articles to be operated upon whether cast or wrought iron, heavy or light-steam superheated up to about 1,000° Fahr. is turned into the muffle, and the formation of magnetic oxide immediately commences. The heated iron or steel decomposes the steam, taking up certain definite proportions of oxygen forming Fe,O, or magnetic oxide. The thickness of the coating depends upon the duration of the operation, and the time varies from six to eighteen hours. Magnetic oxide so formed is in its substance black, but when the articles first come out of the muffle they have a bloom on them very much like that upon a ripe damson.

The Barff process, as it was carried on by the professor, was necessarily an expensive one, because there was required first the steam, then the superheating of it, and the external heating of an iron muffle, which had not a very long life. Wrought requires less heat than cast iron, so that a muffle lasts much longer for wrought alone than if it has to be used for both. A good deal of beautiful work was done during the three or four years the process was worked by the professor, but it was not taken up by the iron trade, on account of its cost.

At one time it looked as though what had promised so well was to turn out commercially unsuccessful or at best to have but a very limited application.

The ingenuity and perseverance of the Bowers, father and son, came to the rescue, and as a result we have the triumph of to-day as is shown in the samples exhibited.

Mr. A. S. Bower in his Boston paper gives a very interesting account of the many difficulties with which he had to contend before arriving at satisfactory results, and then proceeds to give an account of the present method of treatment and the chemistry of the process.

There was commenced a series of experiments with carbonic acid chemically produced by the decomposition of chalk, the idea being that three equivalents of iron would unite with four of carbonic acid, forming one equivalent of magnetic oxide and four of

FIG. 1.-LONGITUDINAL SECTION OF BOWER FURNACE ALONG LINE
1-2 IN FIG. 3.

carbonic oxide, if the heat were sufficiently high. This reaction is expressed symbolically thus: (1) 3 Fe + 4 (CO2) = Fe3O4 + 4 (CO). This is the simplest action that could take place, but it was evident from the results that something quite different was obtained, inasmuch as the coating was very light in color, pleasing to the eye, but easily removed, and in that sense entirely differing from the articles you see before you. This coating, from effects exactly similar and designedly produced by a studied manipulation in the furnaces in successful operation in England, France and here, proves pretty conclusively that carbonic acid, practically pure, produces upon iron, at an elevated temperature, a film which is, in composition, a mixture of Fe 0 and Fe, 04, or, at all events, it is nearer the metallic state than is magnetic oxide. But even supposing that the results obtained by the carbonic acid had been successful as then carried out, the objections referred to concerning the air process would still exist, as external heat and a closed iron

muffle would always be necessary. I therefore proposed to use a fuel-gas producer, similar in principle to the Siemens' generator, but altered practically to suit other requirements, to burn the combustible gases thus produced with a slight excess of air over and above that actually required for perfect combustion, and to heat and oxidize the iron articles, placed in a suitable brick chamber, by these products of combustion. I also arranged a continuous regenerator of fire-clay tubes underneath the furnace, so that the products of combustion leaving the oxidizing chamber passed outside the tubes, imparting a portion of the waste heat to thein, which was taken up by the in-going cold air passing through their interior on its way to the combustion chamber. I had hoped in this way to be able so to regulate the excess of air over that required for complete combustion as to be able to produce magnetic oxide directly, instead of the lower and useless oxide or combination of oxides produced by carbonic acid alone. I obtained some beautiful results, and some again were unaccountably bad, and I soon found that it was as difficult to regulate the precise amount of oxidation as it was first in the Bessemer process, and I was fortunate enough to hit upon an almost parallel remedy—that is to say, I increased the quantity of free oxygen mixed with the products of combustion, and oxidized the iron articles to excess during a fixed period of generally 40 minutes, when magnetic oxide was formed close to the iron and sesquioxide over all. Then for 20 minutes I closed the air inlet entirely, leaving the gas-valve open, and so reduced the outside coating of sesquioxide to magnetic oxide by the reducing action of the combustible gases alone.

The excess of oxygen in the first instance produces Fe, О, or sesquioxide of iron, and the under surface of this, being in contact with metallic iron, undergoes reduction to magnetic oxide in the following manner: Four equivalents of sesquioxide unite with one of metallic iron, forming three equivalents of magnetic oxide, or, symbolically (2) 4 (Fe2 O3) + Fe = 3 (Feg (4).

When deoxidizing by combustible gases, consisting mainly of carbonic oxide, three equivalents of sesquioxide unite with one of carbonic oxide and form two equivalents of magnetic oxide and one of carbonic acid, and, symbolically (3) 3 Fe, Og) + CO = 2 (Fes O + CO2). Another method of reduction is by carbon itself, when the formula stands thus: (4) 3 Fe2O3 + C = 2 (Feg O1) + CO. Formula (3) is also the reaction when rusty iron is reduced by producer gases, and which consist largely of carbonic oxide, and by

the specimens exhibited it will be seen that articles completely pitted with rust may have their surfaces rendered rustless. In this case the periods of oxidizing and deoxidizing are reversed—that is to say, the latter occupies 40 and the former 20 minutes. No oxidizing is theoretically necessary, but practically a certain amount is requisite to keep up the heat in the chamber, which, of course

FIG. 2.-TRANSVERSE VERTICAL SECTION ALONG LINE 3-4 IN FIG. 3.

could not be done unless combustion took place some time or other. I only mention the reduction by carbon as exemplified by Formula No. 4 because, while experimenting with a furnace, I was asked by the proprietors of a valuable red-oxide deposit, which was found in so finely divided a state as to be capable of being used at once as a paint, whether I could reduce it to a magnetic oxide. I tried to do so by carbonic oxide, but I found that only the surface of it was affected, and that even this, when taken out of the furnace, speedily returned to its original red color by the combined actions of the hot unconverted material underneath and the air above. It will be found from formula (4) that 2 pounds of carbon are required to reduce 100 pounds of red oxide. This I mixed intimately, in the shape of powder, with the red oxide, brought the mixture to a red heat and