1900] on Experiments at High Temperatures and Pressures. 559 



can be brought about by temperatures of the order of 2000°C., and 

 pressures of more than 50 and less than 100 tons per square inch. 

 There is some uncertainty, as already mentioned, in regard to the 

 actual pressures operative during the trials. Prof. Tammann has, 

 however, obligingly drawn my attention to the fact that the equi- 

 librium curve graphite-diamond may nevertheless have been crossed, 

 but that no diamond was formed because time for crystallisation was 

 not allowed under the conditions of the experiment. I confess my 

 idea in making the trials was that the amorphous carbon or graphite 

 might be forced to melt, and then that the conditions would require 

 it to recrystallise as diamond — not, of course, in the form of large 

 clear crystals, but rather in the style of bort or black diamond. It 

 is, however, true that the pressures used may have prevented any 

 melting at all, and that it may have been a question of recrystallisa- 

 tion of graphite, in which case the addition a.h initio of diamond 

 crystals would, as suggested Ijy Prof. Tammann, have been of advan- 

 tage in promoting crystallisation in that form. 



The experiments described have only been rendered possible by 

 the invention of high speed steel, which keeps its hardness up to 

 nearly or quite a red heat, and any further advance — mainly in the 

 direction of the allowance of more time — must wait for improvements 

 in that material. It may very well be, however, that the limits of 

 temperature within which crystallisation in diamond form can take 

 place, are really very narrow at any pressure — and in this case it will 

 be a matter of very great difficulty to make an apparatus in 

 which the conditions could be kept constant for a sufficient length 

 of time— and the difficulty would be greater the higher the 

 temperature. 



It is noteworthy from this point of view that in Moissan's artificial 

 production of diamond very much lower pressures and temperatures 

 were used than those just described. I have shown* that, using iron 

 as a solvent, it is highly improbable that Moissan attained a pressure 

 of more than 20 tons/sq. inch, and when silver was employed the pres- 

 sure would be much lower. A similar criticism places the effective 

 temperature of formation of diamond in iron or silver spheroids at 

 something of the order of 1500° C. Comparing the experiments of 

 Moissan with those described above, it looks as if Roozeboom's opinion 

 is at present the most probable — viz. that solvents are necessary in 

 order to depress the crystallisation point of diamond to a temperature 

 at which the transformation to graphite is slow enough for rapid 

 coohng to interrupt it. In this case the next step would be to repeat 

 the experiments I have described at the highest possible pressure in 

 the presence of iron, though Mr. Parsonsf has already made some 

 trials in this direction with negative results. We have, however, many 



* Journ. Ghem. Soc, xciii. 1908, 1351. t Loc. cit. 



