METALS. 



427 



by the author is to introduce the residue, and 

 filter into a porcelain crucible, having at the 

 bottom three parts of nitrate of potash and 

 one part of hydrate of potash ; heat to fusion, 

 dissolve and precipitate with chloride of bari- 

 um. The phosphorus and silica will be con- 

 tained in the same solution, and can be deter- 

 mined separately. Jour. Franklin Institute. 



How to t determine the Carbon chemically 

 combined with Iron. The American Journal 

 of Science for May contains a translation of a 

 paper presented to the Chemical Society of 

 Paris, by Prof. Eggertz, Director of the School 

 of Mines, Falens, Sweden, on a new process 

 for determining the amount of carbon chemi- 

 cally combined with iron. We quote : 



When steel or pig-iron containing carbon in chemi- 

 cal combination is dissolved in nitric acid, a soluble 

 brown ^coloring matter is formed whose coloring 

 power is quite intense, and the solution assumes a 

 tint which is dark in proportion to the quantity of 

 the chemically-combined carbon. 



Iron and graphite (or free carbon) do not influence 

 this coloring ; for the solution of nitrate of iron is 

 colorless, or at most slightly greenish, unless ex- 

 tremely concentrated, and graphite is insoluble in 

 nitric acid. 



Thus in dissolving two pieces of different steels of 

 the same weight in nitric acid, taking care to dilute 

 the darker solution until the two liquids present ex- 

 actly the same color, it is very evident that the more 

 highly carburetted steel will furnish the larger quan- 

 tity of liquid, and the proportion of the volumes will 

 indicate the relative proportion of color in the two 

 steels. 



If, now, the composition and content of carbon of 

 one of the steels is known, the absolute percentage 

 of carbon in the other steel may be immediately 

 deduced. 



Suppose that 1 gram, of each of two steels (a and J) 

 have been dissolved, and that the volumes of the 

 two solutions brought to the same degree of coloring 

 bear the relation to each other a : b : : 5 : 7. Know- 

 ing that the steel (a) contains 1 per cent, of car- 

 bon, you at once deduce that the steel (b) contains 

 1.4 per cent, of carbon. 



In applying this method of analysis, certain pre- 

 cautions must be taken, which we proceed briefly to 

 point out. 



In a cylindrical test-tube dissolve gradually in the 

 cold 10 centigrams of wrought iron, steel, or cast 

 iron reduced to a fine powder, in H to 5 cubic centi- 

 metres of nitric acid of 1.2 specific gravity (about 25 

 Baume). ^The use of nitric acid containing hydro- 

 chloric acid must be avoided, because the solution of 

 iron would have a yellow tint. 



In proportion as the metal contains more carbon, 

 more nitric acid must be used. After some time 

 when the chief part of the metal appears to be at- 

 tacked, place the tube in a water-bath to the depth of 

 about fifteen milli metres, and warm it to 80 Centi- 

 grade. In this position only the lower part of the 

 tube is m contact with the warm water : a movement 

 takes place m the acid which favors its reaction upon 

 the metal a slight disengagement of carbonic acid 

 from all the particles of carbon maybe observed. 

 The operation should always be conducted under the 

 same circumstances as to heat and length of time. 



The evolution of gas having ceased (in operating 

 upon steel the reaction must continue two to three 

 hours), place the tube in a large vase filled with water, 

 to bring the solution always to the same temperature. 

 This precaution is indispensable, because the same 

 liquid is darker when warm than when cold. After- 

 ward, pour off as exactly as possible the clear liquid 

 into a graduated burette. Upon the black residue 



remaining m the tube pour some drops of nitric acid, 

 and heat carefully over a lamp. If there is no fur- 

 ther liberation of gas, the residue consists of nothing 

 but graphite or silica. Cool the new solution, and 

 add it to that which is already in the burette. ' 



The liquid is then diluted with water until its 

 color corresponds exactly with that of the normal 

 liquid, which latter should be of such a degree of 

 concentration that each cubic centimetre represents 

 .0001 gram of carbon. 



If, for instance, this normal liquid is prepared from 

 cast steel containing exactly T 8 <nj of one per cent, of 

 carbon, one decigram of that steel must be dissolved 

 in 8.5 cubic centimetres of nitric acid ; 100 grams of 

 steel containing 85 centigrams of carbon would thus 

 be dissolved in 8 ; 500 cubic centimetres of the normal 

 solution, 100 cubic centimetres of that solution would 

 represent one centigram of carbon, and consequently 

 one cubic centimetre of the normal solution would 

 represent .0001 gram of carbon. 



The normal solution does not keep, and should be 

 often renewed, since it becomes perceptibly paler 

 even within twenty-four hours. For it, however, 

 may be substituted a dilute alcoholic solution of 

 sugar, properly caramelized, brought to exactly the 

 same tint ; this solution keeps much longer without 

 sensible change. 



As one gram of iron cannot readily be dissolved in 

 less than 15 cubic centimetres of nitric acid, it follows 

 that a proportion of carbon less than T Vg of one per 

 cent.* cannot be estimated by means of the normal 

 liquid, but this minimum is seldom found in prac- 

 tice. 



If the proportion of carbon exceeds ^ of one per 

 cent., the ferruginous solution is so concentrated that 

 it has a light greenish tint, which renders its com- 

 parison with the normal liquid difficult. It that case 

 a normal liquid of one-third the strength is prepared by 

 diluting the normal liquid with twice its volume of 

 water; then each cubic centimetre of the liquid repre- 

 sents only one-third of the ten-thousandth part of a 

 gram of carbon. When the proportion of carbon in 

 the specimen to be analyzed is very large (as, for 

 instance, in white cast iron), only .05 gram of the 

 metal must be taken for analysis, and in that case 

 half a cubic centimetre of its solution corresponds to 

 a cubic centimetre of the normal solution. If the 

 metal to be analyzed contains graphite, the latter 

 must be collected on a filter before the solution is put 

 into the burette. 



This method is more exact in proportion as the 

 percentage of carbon is smaller. With an accurate 

 balance, and with suitable arrangements, a great 

 number of determinations of carbon close enough 

 for practice can be effected in a time relatively very 



All the Bessemer steel made at Edskin, in Sweden, 

 is marked after hammering (apres I'etirage) by figures 

 expressing its hardness as ascertained by this color- 

 measuring analysis of Prof. Eggertz. 



It is obvious that only burettes of perfectly color- 

 less glass must be used, or at least they must all have 

 exactly the same tint. 



Chemical Nature of Cast Iron. A committee 

 of the British Association, appointed to investi- 

 gate this subject, reported that they have been 

 enabled to prepare iron of such purity, that 

 it contained in 100 parts only 0.00025* parts 

 of sulphur, and was entirely free from phos- 

 phorus and silicon. The amount of the iron 

 taken for each analysis was about thirty 

 grammes. The committee expressed a hope 

 that in another year a great deal of very 



* O. D. Allen, of the Freedom Iron-works, Penn., has 

 found it quite practicable by a modification of thia 

 method to distinguish between irons containing respec- 

 tively ^% and T Vo of one per cent, of carbon. 



