870

STEEL.— Bessemer Process. Gases Evolved.

10-46 volumes, leaving 12-15 volumes (13-42 parts byweight) of oxygen to combine with other elements.This will require 11-74 parts silicon, while thecarbon going off as gas amounts to 5-27 parts. Car-bon is, therefore, now being oxidised more rapidlyin proportion to the silicon than was the case whenNo. 1 sample was taken.

Sample 4. This was taken ten minutes from thestart, and as soon as possible after the boil had com-menced. The complete spectrum was now visibleand constant, and remained so to the end of theblow. The gas contained—

Carbonic acid,. 3'58

Oxygen. none.

Carbonic oxide,. 19*59

Hydrogen (not estimated),. 2*00 probably

Nitrogen,. 74'83

100-00

The great increase in the proportion of carbonicoxide, and corresponding decrease of carbonic acid,will here be noticed. This accounts for the increasedluminosity of the flame at this period, as we havenow a large volume of gas actually burning at themouth of the convertor. Assuming all the oxygennow being evolved in combination to be deriveddirectly from the blast at present passing throughthe iron, we have for the 74-83 volumes nitrogen,19-84 volumes oxygen entering the convertor; but thecarbonic acid contains 3-58 volumes, and the carbonicoxide 9-79 volumes; together = 13-37 volumes incombination with carbon, leaving 6-47 volumes forother work. This would oxidise 6-25 parts silicon,and there are 9'6 parts of carbon being burnt, so thatthe conditions are now reversed, carbon being oxi-dized at a more rapid rate than silicon. This is thepoint at which the carbon lines become well defined,and the analysis, we think, shows why this is so.

No. 5 Sample, taken twelve minutes from start,gave—

By Absorption,

By Explosion.

Carbonic acid,.

2-47

23

Oxygen,.

absent.

—

Hydrocarbons,.

absent.

—

Carbonic oxide,.

29'58

29-3

Hydrogen,.

—

2-16

Nitrogen,.

67-95

66-24

100-00

100*00

This sample of gas was specially tested for hydro-carbons, as being the most likely period of the blowto detect them, if they were formed to any extent.But although the gas was left in contact with fumingsulphuric acid for twelve hours, no absorption tookplace. As the results, by explosion, agree in showingthe absence of these gases, I think we may fairlyconclude that none are formed.

Sample No. 6 was taken fourteen minutes fromstart, and four minutes from finish of blow. It gave—

Carbonic acid,. 1'34

Oxygen,.absent.

Carbonic oxide,. 31*11

Hydrogen (not estimated),. 2’00 probably

Nitrogen. 65*55

100-00

The large percentage of carbonic oxide, and almost [absence of carbonic acid, is remarkable. 65-55 parts |nitrogen correspond to 17-37 oxygen. 16-89 volumes(= 18-66 parts by weight) are combined with 13-45parts of carbon, leaving -53 part of oxygen, whichwould combine with -46 part silicon; so that at thispoint it is practically carbon alone that is beingoxidised, the last traces of silicon disappearing verygradually.

Looking through these analyses, we note the broadfact that carbonic acid is formed in the first part ofthe blow with little or no carbonic oxide, whileduring the latter part carbonic oxide is formed withonly traces of carbonic acid. The question “ Whyis this ? ” suggests itself. Why should the carbontake up twice as much oxygen at one time as anotherunder apparently the same conditions? My ownimpression is, that both this chemical reaction, andalso the difference between this carbon spectrum andthat from carbonic oxide, under ordinary circum-stances, are functions of temperature. It is certainthat, at the commencement of the blow, the tem-perature cannot be much above a yellow heat, whileat the end of the blow it is undoubtedly a goodwhite heat. Now, it appears to me that, under theconditions in the convertor, carbonic oxide is themost stable body at a high temperature, and carbonicacid at a low one. This agrees with the experimentsof I. Lowthian Bell , where he shows that theconditions of equilibrium for these gases in thepresence of metallic iron are—

FuU red ^ " '! ‘!' * 1 |? vo ]; carboni « acid For each 100Approaching whiteness, 11 “ “ j carbonic oxide.

It also agrees with the general observation that, ifthe iron is cold when it comes from the meltingfurnace, the period of the boil is delayed; while, ifthe iron has been kept in a reverberatory meltingfurnace for a considerable time at a high temperature,this period is reached more quickly. The action ofthe oxygen upon the silicon may have something todo with the degree of oxidation of the carbon, for Ifind that the time when the boil takes place dependsclosely upon the quantity of silicon contained in theiron. Thus, with pig containing about 2£ per cent,silicon, the boil takes place about the middle of theblow, while with Swedish iron, which contains aboutonly 1 per cent, silicon, this point is, I believe,reached much sooner.

It had been suggested by N. Williams (see Nature ,March 16, 1871), that the great brightness of theflame might be due to hydrocarbons; but analysisshows that this supposition is untenable. I thinkit more likely that the reason why we get such abright flame from carbonic oxide during the blow is,that it is burning at the mouth of the converter at amuch higher temperature than we get when we ignitea jet of the cold gas, for it is evident that the car-bonic oxide formed in the metal bath must haveeven a higher temperature than the metal itself, andthis is superadded to the temperature developed bythe carbonic oxide burning in air. This is in accord-ance with the explanation given by II. St. Claire