100 BULLETIN 184, imiTED STATES NATI03S1AL MXJSEIUM 



Though a considerable length of time is required for coloring car- 

 bide (cementite) in artificial irons, the action on cohenite is rapid. 

 In some cases the author has found it colored after one minute in 

 the boiling solution. 



Electrolytic etching with this solution has been found by the 

 author to be a very quick and effective method for identifying cohen- 

 ite. For small samples a beaker is convenient, the cathode being 

 the sample; the anode may conveniently be a piece of stainless steel. 

 The current, of about 6 volts and l}i amperes, may be obtained from 

 three dry cells if desired. 



The solution, prepared cold, becomes warm and turns rapidly 

 to orange, red, dark red, and brown. The action on cohenite appar- 

 ently is confined to the orange and red stages, with little additional 

 effect from longer application. The action is very quick, 5 or 10 

 seconds being usually enough to darken cohenite inclusions. It 

 is best to immerse the sample in the hydroxide solution before add- 

 ing the picric acid. The result is not a true etch, but only a color- 

 ation from the deposition of a stain wliich is very easily scratched 

 and is readily removed from the sample by a light application of 

 rouge. Cohenite inclusions not infrequently take on brilliant pris- 

 matic colors which apparently are permanent. 



The action of the electrolytic process is irregular, depending on 

 the volume of current acting upon the surface of the specimen, which 

 in turn depends upon its size. The results, at their best, are very 

 perfect. 



Another method, which the author has used to some extent, is 

 the use of Murikami's reagent. It is used to identify the ternary 

 iron-tungsten carbides in tungsten steels, which are quickly colored 

 blue at room temperature, though cementite is not affected unless 

 the reagent is used boiling. Cohenite is colored brown when the 

 solution is used boiling for 5 to 10 seconds. Schreibersite is affected 

 only slightly, if at all, if the application is brief. 



The solution should be freshly prepared, in the proportions of 

 10 grams of potassium ferricyanide and 10 grams of potassium 

 hydroxide in 100 cc. of water. Like sodium picrate, it is not strictly 

 an etchant but deposits a stain. 



This method has one advantage over electrolytic etching, in that 

 its action is regular and more controllable. On the other hand, the 

 coloration is less satisfactory and scratches appear conspicuously. 

 Even with the finest abrasives, it was found impossible to produce 

 a surface that would not show scratches. 



Examples of all the methods described are shown in plate 63. 



Electrolytic etching with chromic acid is often used to identify 

 carbide in artificial irons, but the above methods have been found 



