NOVF^BER l8, 1909] 



NA TURE 



87 



hydraulic lifting jack— the one I have used will lift fifty 

 or sixty tons — the bore of the hydraulic cylinder being 

 about 4* inches. In order to operate at a high tempera- 

 ture it "is necessary to line the cylinder with some re- 

 fractory substance, and I have generally used magnesia 

 for this purpose, though zirconia or thoria might be better. 

 Purified magnesia is first melted in an electric furnace, 

 and then ground in an iron mortar until it is very fine. 

 The powder is freed from iron as well as possible by a 

 strong magnet, and after being sifted is pressed into the 

 cylinder little by little by hydraulic pressure so as to form 

 a solid plug. This is then bored out with a hard steel 

 drill to the required diameter. In pressing magnesia I 

 have found that it is not possible to thoroughly consolidate 

 the powder in greater thickness than a few millimetres, 

 even under a pressure of 50 tons per square inch. In 

 fact, magnesia is a substance which appears to be almost 

 devoid of the fluid properties so marked in graphite — an 

 essential condition for its use in the apparatus. I have 

 tried various other linings, ground flint, alumina, &c., but 

 they have no advantage over magnesia, and are even more 

 difficult to drill out. .Mumina prepared from the crystal- 

 line hydroxide is very easily compressed into cakes, and 

 makes a good lining, but it is too fusible for experiments 

 on carbon, and is probably more easily reduced. The 



n-nltTTi^ 



n 



Fic. 7. 



cylinder having been lined, the bottom is filled in with 

 Acheson graphite in electrical communication with the base 

 of the apparatus. The substance to be operated upon is 

 placed in the narrow part of the bore, and packed in with 

 graphite or lead if that is suitable. The pressure is applied 

 by a ram of hardened high-speed steel working upon a 

 reservoir of graphite or lead contained in the plug closing 

 the cylinder at the top and electrically cojmected to the 

 other terminal of the supply. The chief uncertainty in 

 regard to the pressure which actually reaches the subject 

 of the experiment lies in the possibility of the ram being 

 held to some extent bv friction against the sides of the 

 C3'lindrical hole in which it works, and in the consolida- 

 tion of the graphite, with reduced fluidity, before it actually 

 flows. One has to trust either to the hardness of the 

 ram or to leave a space round it sufficient to allow graphite 

 to escape, when the apparatus follows the lines of 

 .\magat's standard pressure gauge, but the duration of 

 the experiment is curtailed by the exhaustion of the 

 graphite supply. A correction has to be applied for the 

 pressure absorbed by the lead or graphite in accordance 

 with the results of the preliminary trial. It is fair to say 

 that no tendency of the ram to stick has ever been noticed 

 — on the contrary, changes of volume brought about by 

 heating have made themselves evident at once on the 

 pressure gauge of the hydraulic press. 

 NO. 2090, VOL. 82] 



When working with any form of carbon there has been 

 no trouble in arranging to heat the body which is being 

 compressed by electrical means. It has been found most 

 convenient to adjust the current to about the value re- 

 quired by means of a resistance — large compared with that 

 of the pressure vessel — the latter being short-circuited 

 meanwhile. In making an experiment, the hydraulic 

 press is worked until the desired pressure is attained, and 

 then by opening the switch the current is thrown on to 

 the apparatus. When the magnesia lining begins to melt, 

 the pressure, as shown by the pump gauge, is seen to 

 fall, graphite flows into the magnesia tube, and the pump 

 is worked so as to compensate for this. Under these con- 

 ditions the pressure is probably transmitted without 

 appreciable loss, as the narrow part of the cylinder is 

 now in a fluid bath. After a sufficient time has been 

 allowed the switch is closed, and the pressure kept up by 

 pumping until the apparatus is cold. Originally an 

 apparatus with a cylinder made in one piece was employed, 

 and in this case there was a considerable voltage between 

 the graphite entering the apparatus and the steel walls 

 of the pressure vessel. After a few seconds of intense 

 heating it frequently happened that an explosion took place, 

 due (as could be seen by subsequent examination) to fila- 

 ments of graphite being driven through the magnesia and 



^ 



-iffi,' 



imr 101" 



producing short circuits against the steel vessel. With the 

 construction above described these explosions do not occur, 

 and there is the additional and very real advantage that 

 when an experiment is over the apparatus can be opened 

 in the middle and everything exposed to view. 



A large number of experiments were made on different 

 kinds of carbon and graphite. The weight of material in 

 the highly heated part was generally from i to 2 grams, 

 and the energy supply was at a rate ol 5 to 10 kilowatts 

 for from three to six seconds. The pressure in a successful 

 experiment lay at from 50 to 100 tons per square inch 

 throughout. The magnesia lining was usually melted for 

 a distance up to i centimetre round c,,^ graphite. Now 

 magnesia melts at ordinary pressures at about 2000° C, 

 hut the energy supply is sulTicient to render it possible 

 that temperatures of from 3000° to 4000° C. may have 

 been reached ; it is possible that about 3000° C. was 

 actually attained at the centre of the charge. The results 

 obtained were uniform. No matter what form of carbon 

 (excluding diamond, which was not tried) was packed 

 originallv in the apparatus, the final product was soft, 

 well-crystallised graphite, which agrees with some results 

 of similar experiments described by Mr. Parsons,' but not 

 with the results claimed by Dr. Ludwig.' 



In several experiments the crystalline mass of graphite 

 1 Proc. R.S., 79. 2 Zeits. f ■""■"■" 1902. 273 



