February ig, 1920] 



NATURE 



679 



apparatus and experiments designed to melt graphite 

 under high pressure, his inference tlien being that 

 under pressures up to loo tons per square inch carbon 

 does not follow the same law as many other sub- 

 stances, and does not crystallise as diamond on 

 cooling. 



An interesting discovery was made by Bridgeman 

 in 191 1 when studying the compressibility of mer- 

 cury. He found that it had a remarkable power of 

 penetrating steel containers, a power not possessed 

 by oil or water, which caused them to hurst at 

 niuch lower pressures than when they were charged 

 with oil or water. The phenomenon he attributed to 

 the fact that mercury has the power of dissolving 

 small percentages of iron, and will amalgamate with 

 it when the surfaces are absolutely free from oxide. 



In 1912 Bridgeman published his remarkable re- 

 searches on water under pressures up to 20,000 atmo- 

 spheres. He found that there are four allotropic 

 forms of ice besides ordinary ice, which are found 

 under various conditions of pressure and temperature 

 with determinate regions of stability. All these forms 

 except ordinary ice are more dense than water ; one 

 is remarkable as existing from a temperature of 

 — 18° C. under a pressure of 4500 atmospheres up to 

 a temperature of 67° C. under a pressure of 20,000 

 atmospheres. 



Recently a pressure of from 200 to 1000 atmo- 

 spheres at a temperature between 500° and 700° C. has 

 been applied to compel hydrogen to combine with 

 nitrogen to form ammonia on a great commercial 

 scale, a catalyst being necessary to promote the com- 

 bination and to establish the equilibrium between the 

 gases and their product. This action is reversible as 

 regards temperature and pressure. On the other 

 hand, iron just molten is an energetic catalyst in the 

 transformation of diamond into graphite, but, con- 

 trary to expectations, as we shall see, no amount 

 of pressure that has yet been applied appears to have 

 caused a reversal of this action. 



More than thirty years ago, having suitable 

 apparatus at hand, I made a few experiments to 

 trv the effect of high pressures and temperatures on 

 carbon, compounds of carbon, and some other sub- 

 stances. 



The apparatus consisted of an 80-ton press, under 

 which suitable containers were placed, and a turbo- 

 generator of 24 kilowatts output at 80 volts provided 

 the current. It had been discovered by Cheesborough 

 that the carbon filaments for incandescent lamps 

 became very hard and resilient when heated in a 

 hydrocarbon atmosphere of about 4 mm. absolute pres- 

 sure, and I was anxious to try what would be the 

 result if a rod of carbon were electrically heated when 

 submerged in a liquid hydrocarbon under high pres- 

 sure. Benzine, paraffin, treacle, chloride, and bisul- 

 phide of carbon were tested under a pressure of 

 2200 atmospheres, or about 15 tons per square inch. 

 The results were not successful in producing a hard 

 coating to the rod or in increasing materially its 

 densitv and hardness except in the case of tetra- 

 chloride of carbon, which slightly consolidated and 

 hardened it ; on the contrary, the carbon deposited 

 from the liquids always appeared as soft amorphous 

 carbon like soot. These experiments were extended 

 b\ substituting, instead of the liquids mentioned, 

 silica, alumina, and other substances and increasing 

 the pressure to 30 tons per square inch. When the 

 current density was sufficiently increased the rod was 

 converted to soft graphite. Moissan in 1003 expressed 

 the view that iron in a pasty condition was the 

 matrix of the diamond, and that great pressure was 

 the determining factor, which compelled a minute 

 fraction of the carbon present to appear as diamond; 



NO. 2625, VOL. 104] 



he further refewed to the probability of carbon being 

 liquefied when under a pressure sufficient to prevent 

 its volatilisation, and that from the liquid state it may 

 pass into the crystalline form on cooling. Crookes, 

 in his lecture delivered before the British Association 

 at Kimberley in 1905, emphasised the same view as 

 to the probability of the crystallisation of carbon 

 directly from the molten state on cooling. 



Though my original experiments in 1888 were not 

 favourable to these views, it nevertheless seemed 

 desirable to carry the investigations up to the greatest 

 possible pressures attainable. Experiments were, 

 consequently, resumed in 1907 with a new equipment, 

 which consisted of a 2000-ton hydraulic press and a 

 storage battery of 360 kilowatts output. The battery 

 can be coupled for 2, 4, 8, 16, or 48 volts as 

 required, and the mains and the main switch can 

 carry currents up to 80,000 amperes to the hydraulic 

 press, which is placed by itself in a small, strong 

 house partly below ground, with walls of 2 ft. thick- 

 ness reinforced with steel bars ; the door is of steel 

 3 in. thick, and the roof is of light galvanised iron. 

 The container under the press is further enclosed by 

 2 in. thick telescoping steel rings, raised into position 

 by steel ropes and counter-weights. These pre- 

 cautions, as experience showed, were necessary, as 

 several violent explosions occurred which cracked the 

 steel rings and blew off the roof. A charge of iron 

 and carbon, when confined and raised to a high tem- 

 perature, mav be verv violent if suddenly released by 

 the melting of the poU'-pieces ; also some endothermic 

 compounds have been formed which swelled the con- 

 tainer and allowed the contents to escape. 



Mv experiments confirmed the conclusion at which 

 Threlfall had independently arrived : that under pres- 

 sures up to 100 tons per square inch and very intense 

 heating by electrical current, graphite' is not materially 

 changed. But modifications in the experiments were 

 made and other methods adopted, as will be explained, 

 which in some respects carried the investigation to 

 still higher pressures and temperatures ; these, how- 

 ever, lead to the same conclusion. 



1 propose this evening , to deal chiefly witji the 

 practical or engineering side of the subject, and to 

 review the limits of pressure and temperature which 

 are artificiallv attainable, and to make some com- 

 parison between them and the pressures and tem- 

 peratures occurring in Nature. 



When the blade of a knife is pressed strongly 

 against another blade so as to make a dent in each, 

 the pressure on the boundary surface of the metal 

 at the notch will have averaged from 300 to 350 tons 

 per square inch, according to the hardness of temner 

 of the steel. The pressures on the knife-edges of, a 

 weighing machine when fully loaded are also of the 

 same order. 



When a needle is broken or a oiece of piano-wire 

 is strained to the point of breaking, the maximum 

 tension on the metal will be at the rate of 150 tons 

 per square inch. On the other hand, the pressures 

 that occur in the chambers of large guns do not 

 usually exceed 20 tons per square inch, and the tensile 

 stress on the plates of a ship in heavy weather should 

 not exceed 8 tons oer square inch. From these simple 

 instances some idea is gathered of the limitations 

 imposed bv materials and dimensions upon apparatus 

 for experimenting at high pressures because of the 

 practical difficulty of hardening and tempering steel in 

 larsje masses. 



When dealing with small amounts of material in 

 each experiment the dimensions aflovv of the container 

 and the ram beinsj made of tungsten steel, which can 

 be hardened and tempered throughout, and not only 

 superficially, as in the case of ordinary carbon steel. 



