198 DR. S. W. J. SMITH AND MR. J. GUILD: A THERMOMAGNETIC STUDY OF 



crystals. Carbide must therefore precipitate from the solution on the surfaces b in 

 fig. 10 (i.) until the concentration of that in solution is lowered to the equilibrium value 

 corresponding with the temperature of the material. Similarly the chemical potential 

 of the iron in solution is higher than that in the separated crystals, and precipitation 

 of iron must occur on the surfaces a until the concentration of the iron in solution is 

 lowered to the equilibrium value. 



The solution near the surfaces b thus at once gets weaker in carbide, and that near 

 the surfaces a weaker in iron. The concentration gradients thus established cause 

 the iron to diffuse in one direction and the carbide in the opposite. This diffusion in 

 its turn causes fresh separation of iron on the one side and of carbide on the other. 

 Hence separation and diffusion will occur continuously until crystallisation is 

 complete. 



The transformation of the eutectoid during subsequent re-heating can be treated 

 in a similar manner by supposing solution to begin again where the crystallisation 

 ended. It is not difficult to see how this view of what happens could be extended to 

 meet other cases in which the structure of the eutectoid is different. 



17. Application to the Case in which the Crystallising of the Eutectoid is 



Interrupted. 



The utility of the above picture of the process of crystallisation can be tested 

 by considering what, according to it, should happen if the process were interrupted 

 before completion, and by examining, thermomagnetically, what actually occurs. 



Imagine, therefore, that the temperature of the furnace, in which the steel lies, 

 is raised slowly before crystallisation is complete. 



The first effect is to lower the rate of escape of heat from the steel to the furnace, 

 and therefore to reduce the rate of crystallisation. The diffusion within the solid 

 solution will continue, however, and will cause the boundary films at a and b (fig. 10) 

 to become respectively richer in iron and in carbide than they were. This will make 

 it possible for crystallisation to continue from them at a higher temperature. 



Crystallisation need not cease at once, therefore, and may go on for some time after 

 the temperature of the material has begun to rise. But, unless the diffusion after 

 interrupted cooling is sufficient to remove the differences of concentration within 

 the still-untransformed material, solution must begin again before the true eutectoid 

 temperature is reached. 



Fig. 11 is an example of results obtained during interrupted cooling. In the first 

 case the cooling was not interrupted until the return of magnetisation was nearly 

 complete. In the third it was interrupted soon after the return had begun. 



The curves show that the gain of magnetisation continues until the temperature 

 has risen appreciably above its value when the cooling was arrested. 



They show also that the subsequent loss of magnetisation begins earlier, but 



