Friday, 4 October 2013

Sir Charles Algernon Parsons 1918 - Diamond Synthesis Using a Wildcat .303 Lee-Enfield Rifle

Main Menu - click above

In 1918, Sir Charles Algernon Parsons delivered a lecture before the Royal Society. It concerned the artificial production of diamonds.

Algernon was extremely rich because he'd developed the steam turbine engine, built his sports-boat "The Turbinia," and subsequently revolutionised shipping. He said "We have now made a bit of money and deserve to have some fun." Over several decades he blew 20 grand on attempts to synthesise diamond.

One of the experiments he described to the Royal Society used a very high velocity rifle. He employed John Rigby & Co., manufacturer of fine guns and rifles, to modify a Lee-Enfield .303 so it shot bullets at 5000 feet per second - Mach 4.6 (standard muzzle vel. 2,441 ft./sec). It had a custom breech to withstand a cordite charge 90% greater than in normal cartridges. The bullets were steel, and half the weight of  regular .303 bullets. A real "Wildcat Cartridge."

He fired these steel bullets, at point blank range, into holes drilled in steel blocks. The holes contained graphite, which he hoped would be transformed into diamond by the heat and pressure of the tremendous impacts.

His rifle experiments are briefly mentioned in this rather quaint video - Man Made Diamonds -

The 1918 lecture is below, followed by his previous Royal Society lectures on diamond synthesis, in 1893 and 1907. The section on the high velocity .303 follows his diamond experiments using a
6-bore shotgun to compress exploding acetylene.

Lecture delivered before the Royal Society on April 25th, 1918


In this paper is given an account of experiments on the artificial production of diamond which I commenced in 1887, and carried on intermittently till the commencement of the War, when they were interrupted. Although the account is not as full as I could have wished, yet it is hoped that from the description of such experiments as relate to the salient features, fol­lowed by a summary of their bearings upon the research, and the conclu­sions at which we have arrived, a fair idea may be gathered of this research.

One reason for writing this paper at the present time has been a publica­tion on the same subject by Otto Ruff in Zeitschrift fur Anorganische Chemie, May 25th, 1917, who also referred to the work of Lummer on the apparently molten aspect of the surface of the carbon of the electric arc.

In my paper to the Royal Society in 1888 were described experiments where a carbon rod heated by a current of electricity (Fig. 1) was immersed in liquids at pressures up to 2200 atmospheres, and where the liquids—benzene, paraffin, treacle, chloride and bisulphide of carbon—were found to yield deposits of amorphous carbon.

In my paper of 1907 allusion was made to experiments in liquids at a pressure of 4400 atmospheres, and to the distillation of carbon in carbon monoxide and dioxide at this pressure with similar results, also to an attempt to melt carbon at pressures up to 15,000 atmospheres, which produced soft graphite, and an experiment where a carbon crucible, con­taining iron previously heated and carburised in the electric furnace, was quickly transferred to a steel die, and while molten and during cooling subjected to a pressure of 11,200 atmospheres, the analyses showing less crystalline residue than if the crucible had been cooled in water.

It was also emphasised that the pressure of 11,200 atmospheres must be greater than could be produced in the interior of a spheroidal mass of cast iron when suddenly cooled, and that the inference from these experi­ments was that mechanical pressure is not the cause of the production of diamond in rapidly cooled iron, as had been supposed by Moissan. This conclusion appears to us in the light of our more recent experiments to be one of great importance, and it will be further discussed in this paper.

It may be well to state that, in order to facilitate a clearer view of the bearing of each experiment on the subject, they are not placed always in chronological order. The difficulty of ensuring satisfactory experiments and the elusive character of the analyses must be the excuse for the random character of some of the former. The great majority of the experiments were failures as regards results, but a few have given information that was scarcely anticipated when they were devised.

Several thousand experiments have been made and a much greater number of analyses, generally following the methods of Moissan and Crookes; the more important experiments are described at some length, and in most cases are typical of groups or repetitions of the same experi­ment with small variations.

The selection has been chiefly determined by their bearing on the general trend of the results of our own work and the work of others.

Those who are familiar with analyses for the detection and isolation of minute particles of diamond will know of the tendency of such particles to float, and to become lost in the frequent washings. To diminish the risk of arriving at erroneous conclusions the analyses of the more important experiments have generally been repeated several times.

Experiments under high pressure

In the experiments designed to test chemical reactions under high pressure, where the charge was heated by passing an electric current through a central core (Fig. 2) small residues of diamond occasionally occurred. A review of these experiments, however, indicates in most cases an association with iron, whether introduced intentionally, or present from the melting of the poles, or from other causes.

Experiments designed to melt carbon under pressure by resistance heating

In the attempts to melt carbon under pressure by this method (Fig. 3) heat was applied for a duration of 5 seconds, sufficient in amount to melt the graphite core six times over, with the result of only altering the structure. Richard Threlfall independently came to the conclusion from his experiments at about the same time, 1907, that under 100 tons per square inch, graphite, electrically heated, remained graphite.

It appeared, however, desirable further to investigate the possibility of carbon losing its electrical conductivity when approaching its melting point, as alleged by Ludwig and others, and of thus shunting the current from itself on to the contiguous molten layers of the insulating barrier surrounding it. There had, however, been no indication of this having occurred, even momentarily; the evidence was rather that the graphite core had been vaporised and condensed in the surrounding parts of the charge, yet it was thought well to repeat the experiment with rods of iron and tungsten embedded in the core, so that should the temperature of volatilisation of the metals under a pressure of 12,000 atmospheres exceed that necessary to liquefy carbon under the same pressure, the presence of these metals might produce a different result. No change, however, occurred, though in one experiment the pressure was raised to 15,000 atmospheres.

Experiments designed to melt carbon under pressure by the rapid compression of flame

A different mode of attack was then arranged, which would ensure that carbon should be subjected to an extremely high temperature concurrently with high pressure, obtained by the rapid compression of the hottest possible flame, that of acetylene and oxygen, with a slight excess of the former to provide the carbon.

The arrangement was as follows -

A very light piston made of tool steel was carefully fitted to the barrel of a duck gun of 0.9 inch bore; the piston was flat in front, lightened out behind, and fitted with a cupped copper gas check ring, the cup facing forward; the total travel of the piston was 36 inches. To the muzzle of the gun was fitted a prolongation of the barrel, formed out of a massive steel block, the joint being gastight. The end of the bore in the block was closed by a screwed-in plug made of tempered tool steel, also with a gastight collar. A small copper pin projected from the centre of the plug to give a record of the limit of travel of the piston.

The gun was loaded with 2 drachms of black sporting powder, which amount had been calculated from some preliminary trials. The barrel in front of the piston was filled with acetylene and oxygen, with a small excess of acetylene. It was estimated that this mixture would explode when the piston had travelled about half-way along the bore; when fired the piston travelled to within 1/8 inch of the end, as had been estimated, giving a total compression ratio of 288 to 1.

Result. The surfaces of the end plug, the fore end of the piston, and the circumference of the bore up to 3/8 inch from the end of the plug had been fused to a depth of about 0-01 inch and were glass hard, the surface of the copper pin had been vaporised and copper sprayed over the surface of the end plug and piston.

The end plug showed signs of compression, and the bore of the block for 3/8 inch from the plug was enlarged by 0-023 inch in diameter, both deformations indicating that a pressure of above 15,000 atmospheres had been reached. A little brown carbon was found in the chamber, which was easily destroyed by boiling sulphuric acid and nitre with no residue. There was a small crystalline residue from the melted layer of the end plug, from which was isolated one non-polarising crystal, probably diamond, but too small to identify with absolute certainty.

Considering the light weight of the piston and the short duration of the exposure to heat, also the small diameter and volume of the end clearance space, the observed effects would seem to indicate that a very abnormal temperature had been reached, many times greater than exists in the chambers of large guns. There was, however, no evidence of any melting and recrystallisation of the free carbon present.

Experiments with high velocity bullets

As it seemed desirable to try the effect of still higher pressures, a rifle, 0-303 inch bore, was fitted with a specially strong breech mechanism by Rigby, capable of withstanding a charge of cordite 90 per cent, in excess of the service charge.

The gun (Fig. 6) was fixed in a vertical position on the wall of the armoured press house, with its muzzle 6 inches from a block of steel, in which a hole 0-303 inch diameter had been drilled to a depth somewhat greater than the length of the bullet, and in alignment with the bore of the gun; the trigger was pulled by a string from without. Cylindrical bullets of steel with a copper driving band were used, shorter than the service bullet, and about one-half of the weight, some with cupped noses to entrain material, some with coned noses to match the bottom of the hole in the block. The velocity with 90 per cent, excess charge was estimated to be about 5000 ft. per sec.

The substance to be compressed was placed either at the bottom of the hole when the coned-nose bullet was used, or over the mouth of the hole when the cupped-nose bullets were used. Some of the bullets were of mild steel, but those with cupped noses were of tool steel.

The substances placed in the hole included graphite, sugar carbon, bisulphide of carbon, oils, etc., graphite and sodium nitrate, graphite and fulminate of mercury, finely divided iron and fine carborundum, olivine and graphite. After each shot (Fig. 7) the bullet and surrounding steel were drilled out, and the chips and entrained matter analysed.

Several experiments were also made with a bridge of arc-light carbon just over the hole, raised to the limit of incandescence by an electric current, and the shot fired through into the hole at the moment the carbon commenced to vaporise, as observed in a mirror from without. Also an arc between two carbons was arranged just over the hole (Fig. 8) and the shot fired through it, as also through a crucible of carbon with a very thin bottom containing a little molten highly carburised iron.

Of all these experiments the only ones that yielded a reasonable amount of residue were one made with graphite wrapped in tissue paper, the bullet, however, in this case grazed the side of the hole, thus producing some molten iron by the friction; and the shots through the incandescent bridge, where again some molten metal would probably occur. The residues were in all cases exceedingly small and not more than would be produced from a small amount of iron melted, carburised and quickly cooled. There was no evidence of any incipient transformation of carbon in bulk into diamond that could be detected by analysis.

A bullet was also fired into a long hole, 0-303 inch in diameter, bored in a steel block and filled with acetylene gas, retained by gold-beaters' skin over the mouth, thus repeating the flame experiment (but in this case without oxygen) on a small scale with the intensest pressures available. The residue was nil.

The pressure on impact of a steel bullet fired into a hole in a steel block which it fits is limited by the coefficient of compressibility of the steel, and with a velocity of 5000 ft. per sec. is about 2000 tons per sq. inch. Measurements made from a section through the block and bullet (Fig. 7) showed that the mean retarding force on the frontal face, after impact till the bullet had come to rest, was about 600 tons per sq. inch.

Several experiments were made by substituting a tungsten-steel block, and a hole tapering gently from 0-303 inch at the mouth to 0-125 inch at the bottom, and using a mild steel bullet, which on entry would be deformed and a greatly increased velocity imparted to the nose. Progressively increased charges were used, and even with relatively small charges the block cracked on the second round. With the 90 per cent, excess charge, the block always split on the first shot, but this probably occurred after impact, and not till the full instantaneous pressure had been exerted, which was estimated to be greater than with the plain hole, probably over 5000 tons.

Only graphite was placed at the bottom of the hole in these latter experi­ments, and the analysis yielded nothing.

Experiments on pressure in cast iron when cooled

It has been generally assumed that iron rich in carbon expands on setting, and that this supposed property is a contributory cause in the formation of diamond.

Several experiments were made by pouring iron saturated with carbon from the electric furnace through a narrow git into a very massive steel mould, closed at the bottom with a breech screw (Fig. 9). When cold, the breech screw was easily removed, and there was no sign of any appreciable pressure having come on the threads. Not being sure that, because of capillarity, the corners of the mould had been quite filled, a steel mandril was, immediately after pouring, forced down the git-hole by a press giving a fluid pressure in the mould of 75 atmospheres. The observed pressure on the breech screw appeared not to have exceeded this pressure. Highly carburised iron, therefore, does not expand with any considerable force on setting.

The reason why a lump of cast iron thrown into a ladle of molten metal first sinks to the bottom and soon rises and floats on the surface is probably that cast iron is about seven times stronger in compression than in tension.

Therefore when a sufficiently thick layer of the cold metal has been heated the interior is torn asunder by the expansion of the outer skin, and the specific gravity of the whole mass is diminished. (See Mr Wrightson's paper " On Iron and Steel at High Temperatures ", with discussion, Journal of the Iron and Steel Institute, No. 1 for 1880.)

We may therefore safely conclude that when iron is suddenly cooled, the only compressive bulk pressure that is brought to bear on the interior is that arising from the contraction of the outer layers after setting, and with highly carburised iron this can only be small because of the low tensile strength of the metal.

Gases ejected from cast iron on setting

As bearing upon the question of the possibility of the occluded gases playing a part, Moissan was the first to observe that spherules or small spheres of iron with cracks and geodes never contained diamond. We have made experiments by pouring highly carburised iron, alloys and mixtures on to iron plates, the cooling taking place from one side only, and under such conditions no diamond results; in fact it only occurs when the ingot or spherule is cooled on all sides nearly simultaneously, so that an envelope of cold metal is formed all over before the centre sets.

Since my paper in 1907, the experiment of heating iron in a carbon crucible and transferring it to a steel die and subjecting it to 11,200 atmospheres pressure has been repeated, and it has been found that if the iron is allowed to set before the pressure is applied the amount of diamond is much greater than if pressed when very hot and molten, and that it is then about the same as when the crucible is cooled in water. The only reason that suggests itself to account for this is, that when pressure is applied while the iron is very hot some of the latter permeates the carbon of the crucible, and because of the greater specific heat and lesser con-ductility of the carbon, the iron next to and in the carbon remains molten after the ingot has been cooled by direct contact with the steel cup on the face of the plunger. Thus, when cooling, the occluded gases have a free exit from the ingot, through the molten metal (which is pervious to gas) into the carbon of the crucible, and are not retained in the ingot to the same extent as when it is set and enclosed in an envelope of colder iron impermeable to the gases before pressing.

The experiments of Baraduc Muller (Iron and Steel Institute, Carnegie Scholarship Memoirs, 1914, p. 216), on the extraction of gases from molten steel, showed that steel is permeable to gases down to 600° C.

Other experiments

The action of water on carbide of calcium, and of concentrated sulphuric acid on sugar for 6 hours under pressure of 30,000 atmospheres, was tried; in both cases amorphous carbon was formed and no diamond.

Hannay's experiments were repeated, where paraffin and Dippel's oil with the alkali metals, especially potassium, were sealed in steel tubes and subjected to a red heat for several hours. The analysis gave no diamonds; in fact it became apparent that when hydrocarbons or water were relied on to produce pressure, the latter could only exist for a short time at the commencement, for when a red heat was reached the hydrogen escaped through the metal, and the oxygen combined with the steel.

We did not analyse the steel tubes themselves. Many experiments were, however, tried with central heating under the press at 6000 atmospheres, and nothing was obtained of interest with the substances used by Hannay, unless as previously mentioned, some iron was present.

Priedlander's experiment was repeated, where a molten globule of olivine, in a reducing flame, or with carbon added, was stated by him to contain minute diamonds. An experiment was made with molten olivine in a carbon crucible in a wind furnace stirred with a carbon rod, with and without an electric current passing between the rod and crucible.

Many experiments were also tried at 6000 atmospheres under the press with central heating with olivine associated with carbon, hydrocarbons, bisulphide of carbon, water, etc., also with blue ground from Kimberley instead of olivine. The results of the analyses were in all cases negative, except occasionally when metallic iron was present. Thus in some cases the olivine or blue ground was partially smelted by the heating carbon rod or by the associated hydrocarbons, etc., when such were added, and iron globules were formed. In these, diamond was occasionally found when cooling was rapid and they were centrally situated in the charge.

Very quick cooling. To test the action of very quick cooling a carbon crucible of 2 inch internal diameter charged with iron, sugar carbon, 2 per cent, silicide of carbon, well boiled by resistance heating under atmospheric pressure and 2 per cent, of iron sulphide added, was quickly placed on asbestos mill-board resting on a steel table frictionally held in the bore of the 4-inch mould, below being placed 2 lb. of carbon dioxide snow, and the plunger quickly brought down by the press, subjecting the whole to 6000 atmospheres pressure. When taken out the crucible was intact, the contents had divided into a lower portion consisting of a large grained crumbling mass of graphite admixed with granules of very hard iron, in the centre a rounded pillar of white iron equally hard. The cooling seemed to have been unusually rapid.

The experiment was repeated, the crucible being charged with iron, sugar carbon, 5 per cent, manganese, 5 per cent, cobalt, 2 per cent, silicide of carbon, boiled, and 2 per cent, iron sulphide added.

It was also repeated with water instead of carbon dioxide snow. The result of all these experiments was similar to the first. No diamond was found in any part.

An experiment which seemed to give practically instantaneous cooling was as follows: A small carbon crucible containing iron, with traces of silicon, aluminium, calcium, magnesia and sulphur, was floated on a carbon block on a bath of mercury, all contained in a vessel exhausted to 2 mm. absolute. The crucible was heated by an arc from an upper carbon, the holder passing through a stuffing-box. When the crucible was sufficiently hot and the contents carburised, the upper carbon was thrust down, sub­merging the crucible under the mercury; the cooling was almost explosive and instantaneous—the finely divided iron and graphite on analysis yielded no diamond.

Extremely rapid cooling does not, therefore, seem to be a direct cause in the production of diamond.

Experiments at atmospheric pressure

A convenient method of studying the effect of the association of other elements with iron on a small scale uncontaminated by the vapours of a furnace lining suggested itself, and a series of experiments was made as follows: A deep iron dish was packed tightly with Acheson graphite with a slight dimple in the centre to hold the ingot; above, graphite was filled in loosely to a depth of half an inch covering the ingot. An arc was struck by a carbon on to the ingot submerged in the loose graphite. When the iron was well boiled the surrounding graphite with the ingot in it was dug out entire and thrown into a bowl of mercury covered with water.

The results showed that, using ordinary mild steel, no diamond ever occurred on analysis, but that a small percentage of silicon is absolutely essential; small percentages of aluminium, magnesium, calcium, one or all are important; sulphur, manganese, and cobalt increase the yield, nickel appeared to be a disadvantage. An alloy of iron and 10 per cent, man­ganese, 10 per cent, cobalt, and 5 per cent, silicon gave out much gas when cooled slowly, and on quick cooling in water and mercury most of the spherules were burst and shredded.

Finally about 1 to 3 per cent, of the other elements added to iron ap­peared to give the best results and the spherules were not then burst.

An experiment was made by letting the ingot remain in the bed till it had quite set, hard enough to handle with the iron spoon, and then, cooled in water and mercury. It gave a fair diamond residue.

Experiments on the conversion of diamond to graphite

A clear octahedral diamond was placed in a small carbon crucible and packed loosely with Acheson graphite and heated for 10 minutes to about 1400° G. The diamond was coated with a firm layer of graphite.

After two prolonged treatments with fuming nitric acid and potassium chlorate, alternating with boiling sulphuric acid and nitre, the opaque coating was removed and there remained a blackish translucent skin. When fractured the interior was unaltered and perfectly transparent.

A piece of bort somewhat laminated, after the same treatment, showed the laminations separated by cracks starting from the outside. Upon breaking, the interior surface of the fissures showed an incipient change to graphite, but less rapid than on the outside surface. There was a sinuous pitting, deepest near the outside and diminishing inwards. The substance of the bort between the fissures was unaltered.

The change of diamond to graphite under the conditions described is gradual, the surrounding gases, carbon monoxide, carbon dioxide, nitrogen., hydrogen, and also vapour of iron (as an impurity in the graphite) singly, or collectively, probably play a part, and further investigation as to this seems to be desirable.

Sir James Dewar, in 1880, heated a diamond in a carbon tube to a temperature of 2000° C, while a flow of pure hydrogen was maintained through the tube. The diamond soon became covered with a coating of graphite (Proceedings of the Royal Institution).

A clear diamond plunged into molten iron saturated with carbon at about 1400° C. for 5 minutes was deeply pitted. When removed from the iron small globules of iron adhered to the surface and the pits appeared to occur at these spots.

A clear diamond was disintegrated by cathode rays, the temperature by pyrometer being 1890° C, the splinters were quite black and opaque, but after several prolonged treatments with fuming nitric acid and potassium chlorate, alternating with boiling sulphuric acid and nitre, the coating that remained was a dusky grey, but semi-transparent, the gas present being chiefly hydrogen. (Paper by Parsons and Swinton, January 16th, 1908, Roy. Soc. Proc. A, Vol. lxxx.)

In this latter experiment the surface action appeared to be much less in proportion to the incipient change of the under layer to graphite, and the impression is that at 1890° C. the temperature of bulk transformation is being approached, also that carbon monoxide, carbon dioxide, nitrogen, hydrogen, and iron, one or more, act as catalysts in the change of diamond to graphite.

Experiments on the oxidation of alloys of iron when molten

Iron was melted in a carbon crucible and highly carburised; when it had somewhat cooled, the other elements were added, in small percentages of aluminium, silicon, calcium, magnesium, manganese, iron sulphide, col­lectively and in some cases singly; the crucible was then removed from the furnace and superheated steam blown through a carbon tube into the metal; energetic action took place and much heat was evolved; on analysis, after destroying the graphite, a bulky transparent crystalline residue remained.

With aluminium alone the crystals were chiefly crystallised alumina, and with the other elements the spinels and other crystals were produced; all were transparent and colourless, but when chromium was added some rounded crystals occurred resembling pyrope. When submitted to sulphur dioxide and carbon dioxide the result was the same, but less residue was produced. Under the microscope there appeared to be a small proportion of very small crystals like diamond; these burnt in oxygen. When the bulky residue was placed in a test tube with the double nitrate of silver and thallium, and the density adjusted so that a diamond floated midway between the top and bottom, there collected into its immediate neigh­bourhood after a time an amount of the small crystals which was estimated to be about 5 per cent, of the total residue.

One prolonged treatment of hydrofluoric acid had no apparent effect on the bulky residue, and it required so many treatments to destroy it that we failed to isolate the very small particles whose size did not exceed 1/20 mm.; they were probably lost by flotation. These experiments were repeated many times with the same result, but they merit further investiga­tion, with steam under high pressure and conditions favourable to the formation of larger crystals.

Note. Marsden observed in silver the association of black diamond with crystalline alumina, silicide of carbon, etc., Roy. Soc. Proc. 1880.

Experiments in vacuo

The presence of diamond in some meteorites suggested a series of experi­ments under various degrees of vacuum up to the highest obtainable (Also an impression suggested itself in 1907 that hydrogen had an adverse effect on the formation of diamond).

It is probable that some meteoric matter may have been melted by collision or ejected into space in a molten state and cooled by radiation, and that under such conditions the absence, or diminution, of occluded gases might be a factor conducive to the crystallisation of carbon.

One of the 4-inch diameter pressure moulds (Fig. 10) was used in a preliminary experiment as the container. The crucible was turned out of a 1-1/2-inch carbon rod, and so formed on a stem that the electric current heated the bottom and sides equally. The cover was similarly formed and its holder was electrically connected with the container, but free to move vertically and to rest its weight on the crucible, electrical connection to the container being made by a layer of brass or iron turnings resting on the holder. A current of 1000 amperes at 16 volts sufficed, and the tempera­ture was observed through a glass window at the side of the container.

The crucible was charged with reduced iron and lampblack. The Geryk pump evacuated the container to 1 inch mercury absolute; current was turned on for 15 seconds, the vacuum fell to 3 inches, when it had risen

again to 1 inch current was again turned on. This was repeated three or four times, finally current was applied for 30 seconds and the vacuum again fell to 3 inches. The gas was drawn off and collected, it amounted to a total of 1/2 gallon at atmospheric pressure and consisted of 95 per cent, carbon monoxide, 1 per cent, hydrogen, 2 per cent, hydrocarbon, 2 per cent, nitrogen.

The carbon which formed the crucible and cover contained a large per­centage of silica, but the carbon monoxide was produced chiefly by the action of sand (of which there was a thick layer on the bottom of the container to protect the insulating joint from iron spilled from the crucible) on the carbon of the stem of the crucible. About one half of the iron had been evaporated, and there remained an ingot about the size and shape of a broad bean. It contained rather large graphite crystals and was easily broken. The analysis gave the largest residue of diamond in proportion to the amount of iron of any of our experiments, the largest crystals being 0.7 mm. in length.

This experiment was repeated several times with the same result. The time of cooling of the crucible, from switching off the current to the temperature of setting, was 15 seconds, and probably sufficiently rapid to allow a skin to be formed around the ingot before the centre was solidi­fied, for the configuration of the crucible and cover was such as to ensure nearly equal and simultaneous cooling on all sides of the ingot. At the time, vacuum was erroneously thought to be the chief contributory cause and not the presence of carbon monoxide in large proportion.

High vacuum experiments

The molecular pump not having yet been evolved, a powerful pumping system was arranged, consisting of three steam-jet exhausters in series,

the last ejector of the series discharging into a jet condenser with separate air and water pumps, the former assisted by a steam jet. The two steam-jet exhausters nearest to the exhausted chamber were fed with highly superheated steam at 200 lb. pressure, and the suction pipe to the chamber was 4 inches in diameter; the chamber, 2 feet 6 inches diameter of spherical shape (Fig. 11). A vacuum of 1/6 mm. absolute could be reached. The crucible was placed on a large block of carbon, resting on the base of the chamber, and forming the bottom pole. The cover was insulated from the chamber, and through an oil-sealed gland passed a 2-inch brass rod, carrying a crown holder, with four 2-inch carbons which rested on the lip of the crucible for resistance heating. An observation window was placed at the apex of a long iron cone, projecting from the side of the cover, which gave a good view of the crucible and its contents. The whole of the chamber was submerged in a tank of water, up to the level of the gland in the cover.

Iron and iron alloys were boiled and allowed to cool slowly by radiation, or were rapidly quenched by admitting water through a large valve from the tank into the vacuum vessel. The iron and carbon vapour from the boilings deposited dust and globules on the cover, and on the sides and bottom of the chamber. A very small diamond residue generally resulted from the small iron globules, and also from the dust, but never anything from the ingot remaining in the crucible.

In several experiments water was admitted, which played directly on the crucible, the upper carbons resting on the rim prevented its upsetting by the force of the water, and still there was no residue. In one experiment the carbons were lifted and the charge flowed out, forming spherules of varying size in the water. There was a very small diamond residue from these spherules.

In one experiment a crucible was filled with iron and carbon and closed by a tight carbon cover, a hole bored in the side of the crucible, a massive block of iron placed close opposite the hole and the crucible boiled, the vacuum being under 1 mm. No crystallised residue was found in the deposit on the iron block from this high velocity jet of vapour of iron and carbon.

In another experiment a powerful electromagnet was provided with poles to give a concentrated field, and an arc struck between two carbons was arranged to burn within this field and regulated from without by hand. There was an iron block upon which the arc directed by the field could play and condense its carbon vapour. The analysis gave no diamond.

It was thought that the vapour from boiling iron saturated with carbon might, by the action of bisulphide of carbon, cause a crystalline deposit, but all the experiments to this end yielded no results.

Experiments under X-ray vacuum

Experiments were made under X-ray vacuum in a new chamber of cast iron with very thick walls to absorb the heat, exhausted through an 8-inch diameter suction by a large molecular pump alongside, in series with a dry, high speed, two stage, pump, 12-inch diameter pistons, and last of the series a 3-inch + 2-inch compound Fleuss. The crucible was resistance-heated as before (Fig. 12). No diamond was produced in any of these experiments, except in those where iron, sand, and other elements, with or without sulphur, were first heated and well boiled in the carbon crucible at atmospheric pressure, and after cooling transferred to the vacuum furnace and re-heated by resistance under X-ray vacuum; violent ebullition occurred owing to the liberation of occluded gases, and many iron spherules were ejected, which cooled by radiation and conduction where they fell; diamond was found in these, which burnt in oxygen, but no diamond was ever found in the ingot remaining in the crucible.

It occurred to us to try the effect of great mechanical pressure accom­panied by heat upon small particles and powders, the interstices being exhausted to a high vacuum.

Several experiments were made in the press under a mass pressure of 3000 atmospheres.

A layer of cast-iron turnings resting on a layer of carborundum grit, the exhaustion being effected through a hole in the side of the mould covered by a perforated steel plate within the layer of grit, heat was applied as usual by a central carbon rod.

Analysis yielded some thin crystal plates from the grit which had lain in the line between the cast iron and the suction outlet at the grid, and also from the layer of grit which had lain against the cast-iron turnings which had become heated but not melted by the central carbon rod.

To ascertain the cause of the occurrence of these plates, experiments were made, without bulk pressure, on the concentrated action of the gases given off from cast-iron turnings heated up to a good red, and drawn by a high-vacuum pump through carborundum grit placed in a silica tube heated by a gas burner at the centre of its length to dull red. These yielded similar crystal plates.

Control experiments showed that no similar plates existed in the un­treated grit.

It was also found that the cast-iron turnings would not produce this effect on a second heating unless they had been subjected to CO at atmo­spheric pressure for some hours. Carbon monoxide, sulphur dioxide, cyanogen, hydrogen, nitrogen, oxygen, nitric acid gas, chlorine, ammonia, ammonium oxalate vapour, ammonium chloride, acetylene, or coal gas, produced no plates.

These plates resemble diamond very closely in appearance and form of crystallisation, they do not polarise, and some have triangular markings; they will not, however, burn in oxygen at 900° C, and are completely destroyed by chlorine purified from oxygen and water vapour at 1100° C.; their specific gravity is about 3-2, they are therefore not diamond.

Note. Recent experiments have shown that carbon monoxide passed over molten iron sulphide and then over carborundum grit below red heat at atmospheric pressure also produces these plates, and that if coal gas is substituted for carbon monoxide no plates are formed. Also that only a few of the grains produce plates.

The composition of the grains is

Carborundum ...        ...        ...        36-56
Iron oxide and alumina      ...        .44-09
Lime ... ... ...        ...        ...                10-45
Magnesia ... ...        ...        ...             5-57

Summary of experiments and conclusions

The experiments have shown that all the hydrocarbons, chlorides of carbon, and oxides of carbon tested, deposit amorphous carbon or graphite on a carbon rod electrically heated at any pressure up to 4400 atmospheres, and in a few experiments up to 6000 atmospheres; and that at 15,000 atmospheres carbon and graphite electrically heated are either directly transformed into soft graphite or are first vaporised and then condensed as such.

While the experiment of rapidly compressing a mixture of acetylene and oxygen with the production of temperatures much in excess of that neces­sary to vaporise carbon, accompanied by a momentary pressure of about 15,000 atmospheres, confirms the conclusion that the negative results obtained in the attempts to convert graphite into diamond by electrical heating are not due to lack of temperature; on the other hand, the presence of minute crystals in the molten layer of the steel of the end of the barrel subjected to high gaseous pressures of carbon monoxide, carbon dioxide, and hydrogen appears to be connected with the other experiments bearing upon the inclusion of gases in metal as a factor in the production of diamond.

The experiment of firing a high velocity steel bullet with cupped nose through vaporising carbon into a hole in a block of steel has tested the effect of a momentary pressure of about 300,000 atmospheres on carbon initially near its melting point, and probably raised by adiabatic com­pression by another 1000° C.

The fact that only a very few minute crystals resembling diamond were produced (probably from the iron) raises the question as to whether the duration of the pressure is sufficient to start a transformation of graphite to diamond which can be detected by analysis. We have distinct evidence that, with iron as the matrix, the time is sufficient to form very small crystals which can be identified with some certainty. It therefore seems reasonable to conclude that there was no incipient transformation in bulk, and that however long the pressure of 300,000 atmospheres were applied, it is extremely doubtful if any change would occur.

The pressure of 300,000 atmospheres is between one quarter and one half that obtaining at the centre of the Earth, but vastly greater pressures exist at the centre of the larger stars, and are produced by the collision of large bodies in space; these pressures are many thousands of times greater, and whether they would effect the change it is impossible to predict. On the other hand, a heating effect on large masses of iron might be produced by collisions, and owing to the heat generated by adiabatic compression of the central portions, some of the mass would be melted and subsequently cooled on release of the pressure, so that if heating and cooling under pressure are alone necessary for the production of diamond large stones might result. These considerations, though of interest as bearing upon the presence of diamonds in meteorites and also indicating a possible origin of natural diamond, are of no practical value to us because the pressures required are entirely beyond our reach. There are, however, other considerations arising out of the experiments of Marsden, Moissan, and Crookes, as well as our own, which seem to give some hope of solutions of the problem at issue which lie within the means at our disposal.

A repetition has been made of many of the experiments in which diamond is claimed to have been produced. These have given negative results in all cases except where iron has played a part, as for instance with olivine, when being partly reduced by carbon or a reducing flame, small spherules of iron are produced and may, if the mass is quickly cooled, be found to contain diamond.

The repetition of Moissan's experiments under a variety of conditions and pressures has not only confirmed his results but has thrown, it is hoped, additional light on the causes operating to produce diamond in iron.

The experiments under high pressure in steel moulds, where heating of the charge was effected by a central core through which current was passed, enabled Hannay's experiments with Dippel's oil to be tried under much higher pressures, and more thoroughly than is possible with steel tubes in a furnace.

The Appendix (not reproduced here) gives some indication of the many substances and chemical reactions tested. The results were chiefly negative. The few that were favourable were generally attributable, as has been said, to the presence of iron. It was noticed that the iron seldom contained diamond unless when so situated in the charge as to cause equal cooling on all sides, and it will be remembered that the experiments under atmospheric pressure showed this condition to be essential for the formation of diamond.

In some of the experiments of this group considerable gaseous pressure existed, up to 6000 atmospheres, but it is doubtful if in these the right kind of gas was present or a sufficiency of heating or carburisation of the iron occurred. On the whole, therefore, it would appear that all, or nearly all, the chemical reactions as such, under pressures up to 6000 atmospheres, have given negative results.

The experiments on very rapid cooling would seem to dispel the theory that carbon can be caught in a state of transition, and to lead us to the conclusion that quick cooling is not in itself a cause of the occurrence of diamond in rapidly cooled iron.

Moissan observed that when the spherules of granulated iron were cracked, or contained geodes, no diamond was ever found in them, and he attributed this to want of mechanical pressure. The experiments we have made not only corroborate this fact, but they tend to show, we think conclusively, that the cracks in the spherules act by allowing a free passage for the occluded gases to escape, and the geodes by providing cavities in which the gases can find lodgement without much gaseous pressure occurring in the metal. Further, the experiments have shown that iron when it sets does not expand with appreciable force, and that the only compressive forces that are brought to bear on the interior are those arising from the contraction of the outer layers.

Our experiments further show that when a crucible of molten iron is subjected to pressure more than three times as great as can be produced by these contractile forces, the yield of diamond is not increased. On the other hand, when the conditions of the experiment operate to imprison the occluded gases, then the yield of diamond is about the same as if the crucible had been plunged into water, while if the conditions are such as to allow a free passage through the skin of the ingot, the yield is at once diminished, even though the bulk pressure on the ingot is the same.

The experiment, on compressing acetylene and oxygen, has shown that minute crystals, probably diamond, are produced almost instantaneously in the molten surface of metal exposed on one side to gases consisting of carbon monoxide, carbon dioxide, and hydrogen at very high temperature and at 15,000 atmospheres. Sir William Crookes' experiment described in his lecture before the British Association at Kimberley in 1905 is somewhat analogous; cordite with a little additional carbon was fired in a chamber, the pressure reaching 8000 atmospheres, a few crystals of diamond were found and isolated; this result Crookes attributed to the melting of the carbon under the temperature of explosion and crystallisation under the pressure on cooling.

Under the conditions of the experiment there would be a considerable amount of the surface of the chamber melted and swept into the products of the charge by the turbulence of the explosion, and the spherules of iron would thus be carburised and cooled while still under heavy pressure.

In the acetylene-oxygen experiment there is a molten surface with reducing gases on one side at high pressure, and on the other metal impervious to gases. In Crookes' experiment the globules of metal are surrounded by gases at high pressure. In both cases the metal has solidified with the occluded gases imprisoned by the high external gaseous pressure, for we have seen that the pressure of occluded gases in highly carburised iron when quickly cooled cannot exceed about 1000 atmospheres.

The experiments under vacua from 75 mm. up to X-ray vacua have shown generally that as the vacuum is increased the yield of diamond in the crucible is diminished, and that below 2 mm. none has been detected. But when alloys previously boiled at atmospheric pressure are quickly heated up under high vacuum violent ebullition takes place, from the large volume of gases liberated, and some of the contents are ejected into the vacuum chamber before they have had time and sufficient temperature to part with their occluded gases, and diamond occurs in the spherules so ejected.
The gases occluded in cast iron which are given off when heated in vacuo have been investigated by H. C. Carpenter and others, and the relative amounts of the constituents are found to vary widely according to the previous heat treatment and the nature of the gases in contact with the metal while molten and during cooling; they are carbon monoxide and carbon dioxide, hydrogen and nitrogen.

H. C. Carpenter (Journal of Iron and Steel Institute, 1911) states that, when heating up a bar of cast iron in vacuo in a silica tube, 'After the twenty-fifth heat it was noticed that in the water-cooled areas of the quartz tube a lustrous black ring had formed. On being strongly heated, some of this, evidently carbon, burnt off, leaving a white film, presumably silica. This seems to show that a volatile silico-organic compound, con­taining carbon, hydrogen, and silicon, was evolved from the iron on heating".

It would appear from our experiments that probably a ferro-silicon carbonyl is given off from the iron, for, as has been said, we observed a corrosive action on carborundum by the gas evolved from iron borings at red heat under a high vacuum, and the same action was produced by gaseous ferro-carbonyl, and also by carbon monoxide, previously passed over molten iron sulphide at atmospheric pressure.

Let us consider what happens in an ingot or spherule when rapidly cooled simultaneously on all sides. It is first surrounded by a thin coat of solidified metal which, below 600° C, is impervious to gases. As the coat thickens layer within layer, more and more gas is ejected by the solidifying metal, and its semi-solidified centre, still pervious to gas, receives the charge. As this process progresses the pressure may rise higher and higher, though there may be a limit to the pressure against which the metal is able to eject gas when setting. All we, however, know is, that the mechanical strength of the ingot or spherule places a limit of about 7000 atmospheres on the gaseous pressure, and, as we have already mentioned in the case of some iron alloys, most of the spherules are split or shredded, with an appearance consistent with this view.

Crookes' microscopical examination of diamonds with polarised light supports this view. In his lecture at Kimberley, in 1905, he states: " I have examined many hundred diamond crystals under polarized light, and with few exceptions all show the presence of internal tension.

On rotating the polarizer, the black cross most frequently seen revolves round a particular point in the inside of the crystal; on examining this point with a high power we sometimes see a slight flaw, more rarely a minute cavity. The cavity is filled with gas at enormous pressure, and the strain is set up in the stone by the effort of the gas to escape."

It seems therefore probable, or indeed almost certain, from the accumu­lated evidence, that the chief function of quick cooling in the production of diamond in an ingot or spherule is to bottle up and concentrate into local spots the gases occluded in the metal which, under slow cooling, would partially escape and the remainder become evenly distributed throughout the mass.

As to the condition in which the gases exist within the iron at tempera­tures above 500° C. little is known, though at 200° C. and at 180 atmo­spheres Mond has shown that iron penta-carbonyl is formed. The intimate contact between the occluded gases and other elements, metals or carbides, must favour complex interactions as cooling takes place. Such actions might be concentrated by the heat flow across the metal on quick cooling.

It appears probable that concentration of gaseous pressure causes certain reactions which bring about an association of carbon atoms in the tetra-hedral form—against their natural tendency to assume the more stable form of graphite. It also appears that the conditions may operate to the exclusion of some gas or element inimical to the formation of diamond from certain parts of the metal, viz. the graphite liberated and the cooled metal of the outer layers may absorb some gas or element from the inner portion of the ingot and leave none for the central portion.

The necessity of subjecting the iron to a temperature above 2000° C. before cooling would seem to imply the necessity of carbides of the other metals, such as silicon, magnesium, etc., being present to insure the necessary chemical reactions with the gases at high pressure within the ingot.

In reviewing all our experiments, the greatest percentage of diamond occurred when the atmosphere around the crucible consisted of 95 per cent, carbon monoxide and 1 per cent, hydrogen, 2 per cent, hydrocarbons, 2 per cent, nitrogen, the mean pressure in the vessel being about 1 inch absolute of mercury. The weight of diamond we estimated to be about 1-20,000 of the weight of the iron. If we, for the moment, assume a volume of carbon monoxide at atmospheric pressure equal to 0-69 that of the iron, the weight of carbon contained in it equals that of the diamond.

For the following reasons it would appear that the formation of diamond in rapidly-cooled iron takes place when it is solid or in a plastic condition, or even at a still lower temperature. The rapid pitting of a diamond in highly carbonised iron just above its melting point is so pronounced that the largest diamond hitherto produced artificially would be destroyed in a second or two if the iron matrix were molten. The production of diamond was obtained in an ingot rapidly cooled after it had set sufficiently hard to be handled in a spoon. A similar result was obtained in the case of a crucible placed in the die and subjected to 11,200 atmospheres pressure after the contents had set. Moissan found the diamonds to occur in the centre of the ingots both in the case of iron and also of silver.

It has been seen that iron is permeable to carbon monoxide and hydrogen at temperatures above 600° C, and there appears to be no reason why the concentration of the occluded gases should not take place within the mass as effectively at 600° C. as at higher temperatures, provided that they cannot escape. The most probable temperature, however, may be the point of recalescence at 690° C. These conditions may also operate to exclude some gases from certain portions of the metal.

It would appear that the function of the impervious metal coating thrown around the ingot by quick cooling might be better effected by gas of the same composition as that which the metal ejects on cooling, the pressure being sufficient to ensure that the gaseous pressure around the ingot shall be equal to, or greater than could occur on quick cooling. Such a substitution might result in a larger gaseous content and a larger proportion of the ingot being brought into a suitable condition for the formation of diamond, and the yield might thereby be increased. Some gradations of temperature might still be found necessary to concentrate the reactions. It seems, however, probable that the rate of cooling might be so much prolonged as to obtain much larger crystals and a larger total yield.

The presence of crystals of silica, alumina and magnesia and the spinels, and pyrope associated with diamond in rapidly cooled iron alloys, and also when oxidised by steam and some other gases, appears to have a bearing upon the presence of similar crystals usually found in association with diamond, and to be compatible with the conclusions of Bonney that eclogite is the parent rock of the diamond in South Africa. It seems probable that both the eclogite and the diamond may have been crystal­lised nearly simultaneously from an iron alloy.

Moissan, after a recital of the geological conditions existing in the South African pipes (see Four Electrique, p. 115), came to the conclusion that diamond was not a vein mineral, but must have been evolved in the midst of a plastic mass; and he concludes that iron at high pressure must have been the matrix. Our experiments, however, seem to show that bulk pressure on the metal does not play a part, but that the previous heat treatment, the impurities in the iron and the condition of the gases within the metal, are the important factors.

It is interesting to note that in the best experiments the yield of diamond in rapidly cooled iron has reached 1 : 20,000 of the weight of iron, whereas the weight of diamond obtained from the blue ground of the South African mines is only 1 : 5,400,000. This comparison appears to be confirmed by the relative rarity of microscopic diamonds we have found in the many analyses we have made of blue ground and of the conglomerate from Brazil.

Thus in cooled iron there may be more than 270 times as much diamond as exists in the bulk average of blue ground.



From the Philosophical Magazine for September, 1893Proceedings of the Royal Society.

The primary object of these experiments was to obtain a dense form of carbon which should be more durable than the ordinary carbon when used in arc lamps, and at the same time to obtain a material better suited for the formation of the burners of incandescent lamps.

There were a considerable number of experiments made in which the conditions were somewhat alike, and many were almost repetitions with slightly varying pressures and temperatures. They may, however, be divided into two distinct classes: the first, in which a carbon rod surrounded by a fluid under great pressure is electrically heated by passing a large current through it; the second, in which the liquid is replaced by various substances such as alumina, silica, lime, etc.

The arrangement of the experiment was as follows: A massive cylindrical steel mould of about 3 inches internal diameter and 6 inches high was placed under an hydraulic press; the bottom of the mould was closed by a spigot and asbestos-rubber packing similar to the gas check in guns; the top was closed by a plunger similarly packed; this packing was perfectly tight at all pressures. In the spigot was a centrally bored hole into which the bottom end of the carbon rod to be treated fitted; the top end of the carbon rod was connected electrically to the mould by a copper cap which also helped to support the carbon rod in a central position. The bottom block and spigot were insulated electrically from the mould by asbestos, and the leading wires from the dynamo being connected to the block and mould respectively, the current passed along the carbon rod in the interior of the mould.

The fluid was run in so as to cover the rod completely. The plunger was then free to exert its pressure on the liquid without injuring the carbon. The pressure in the mould was indicated by the gauge on the press.

Experiments. Class I

Among the liquids tested were benzene, paraffin, treacle, chloride and bisulphide of carbon.
The pressures in the mould during the several experiments were maintained at from 5 to 15 tons per square inch; the initial size of the rod was in all cases 1/4 inch, and the current from 100 to 300 amperes.

Results. In some of these experiments a considerable quantity of gas was generated, and the press had to be slightly slacked back during the experiment to accommodate it and maintain the pressure constant.

In all cases there was a soft friable black deposit of considerable thick­ness on the carbon.

In no case was the specific gravity of the carbon rod increased by this process. There was no change in appearance of the fracture, excepting when chloride of carbon had been the fluid; it was greyer in this case.

The rate of burning of samples placed in arc lamps was not diminished by the process. Various rates of deposition were tried, but with the same result; and the conclusion seems to be that under very high pressures, such as from 5 to 15 tons per square inch, the deposit of carbon by heat from hydro-carbons, chloride of carbon, bisulphide of carbon, treacle, etc., is of a sooty nature, and unlike the hard steel-grey deposit from the same liquids or their vapours at atmospheric or lower pressures.

Experiments. Class II

In these experiments the asbestos-rubber packing was omitted, the plunger and spigot being an easy fit in the mould. A layer of coke powder under the plunger formed the top electrical connection with the rod.

No. 1. Silver sand or silica was run around the carbon rod, and pressures of from 5 to 30 tons per square inch applied; the rod was usually about 1/4 inch diameter, and currents up to 300 amperes passed.

Results. The silica was melted to the form of a small hen's egg around the rod. When the current was increased to about 250 amperes the rod became altered to graphite, the greater the heat apparently the softer the graphite. There was no action between the silica and the carbon, the surface of the carbon remained black, and there were no hard particles in or on the carbon rod.

Other substances, such as an hydrated alumina and mixtures of alumina and silica, gave the same results.

The density of the carbon was considerably increased, in some cases from normal at 1-6 to 2-2 and 2-4; in these cases the carbon appeared very dense, much harder than the original carbon, and about as hard as the densest gas-retort carbon. No crystalline structure was visible.

The specimens were treated with solvents, and there appeared no indica­tion of the surrounding substance having penetrated the rod; the carbon was undoubtedly consolidated by 30 per cent.

In some cases, when the material surrounding the rod was alumina saturated with oil, soft crystals of graphite exuded from specimens that had been kept for some weeks.

No. 2. Pure hydrated alumina, carbonate and oxide of magnesia, and lime all rapidly destroyed the carbon rod by combining with it, the hydrated alumina forming large volumes of gas of which it appeared to be a con­stituent. On account of the great diminution of bulk, no analysis was made; the gas issued from the mould explosively at from 10 to 12 tons per square inch. The alumina was found in a crystalline crust, like sugar, around where the rod had been. Hardness that of corundum, almost translucent.

No. 3. The following is the most interesting experiment of the series: On the bottom of the mould was a layer of slaked lime about 1/4 inch thick, over this silver sand 2 inches, then another layer of lime of the same thickness as the former, finally a layer of coke dust, and then the plunger. With a pressure of from 5 to 30 tons per square inch in the mould, and the carbon of from 1/4 to 5/16 inch diameter, currents of from 200 to 300 amperes were passed.

In from 10 to 30 minutes the current was generally interrupted by the breaking or fusing of the rod, or by the action of the lime in dissolving it at the top or bottom. On opening the mould when it had cooled a little, the silica usually appeared to have melted to an egg-shaped mass, and mixed somewhat at the ends with the lime; the surface of the carbon appeared acted on, and sometimes pitted and crystalline in places; silica adhered to the surface, and beneath, when viewed under the microscope, appeared a globular cauliflower-like formation of a yellowish colour, re­sembling some specimens of "bort"' (the bort-like powder is not acted on by hydrofluoric and nitric acids mixed).

After several days' immersion in concentrated hydrofluoric acid, this formation remained partly adherent to the carbon; on the surface of the carbon was a layer or skin about 1/64 of an inch thick of great hardness, on the outside grey, the fracture greyer than the carbon, but having a shining coke-like appearance under the microscope.

The powder scraped off the surface of the rod has great hardness, and will cut rock crystal when applied with a piece of metal faster than emery powder. It has, under the microscope, the appearance of bort, the minute particles seem to cling together; they are not transparent as a rule, and though some such particles are found among them, it is not clear that such are hard.

When a piece of the skin has been rubbed against a diamond or other hard body, the projecting or hard portions have a glossy coke-like ap­pearance.

A piece of the skin will continue to scratch rock crystal for some time without losing its edge. It will scratch ruby, and when rubbed for some time against it will wear grooves or facets upon it. When a cut diamond is rubbed on the surface of the skin, it will cut through into the carbon beneath, making a black line or opening about 1/4 inch long; the facet on the diamond, originally 1/32 inch diameter, will have its corners evenly rounded, and its polished surface reduced to about one half its original area; the appearance of the edges is as if they had been rubbed down by a nearly equally hard substance.

The subject of the last experiment is scarcely sufficiently investigated to warrant any definite conclusions.

The substance in the several ways it has so far been tested seems to possess a hardness of nearly if not quite the first quality. The minuteness of the particles, which appear more or less cemented together, and are less cohesive after the action of acid, make it very difficult to determine their distinctive features.

The mode of formation is not inconsistent with the conditions of pressure, temperature, and the presence of moisture, lime, silica, and other sub­stances as they appear to have existed in the craters or spouts of the Cape Diamond Mines at some epoch.

From the few experiments that have been made it appears that at pressures below 3 tons per square inch, the deposit does not possess the same hardness, though somewhat similar in appearance.

What part the lime and silica play, whether the former only supplies moisture and oxygen which combine with the carbon, or whether the presence of lime is necessary to the action, is not clear.

We may, however, observe that so far it seems as if the lime and moisture combining with the carbon form a gas or liquid at great pressure, which combining with the silica forms some compound of lime, silica, and carbon, or perhaps pure carbon only, of great hardness.

Diamond manufacture

With a view of ascertaining the behaviour of carbon at high tempera­tures and pressures, and in contact with a variety of substances, the above experiments were described in a paper to the Royal Society, June 13th, 1888.

These experiments are of interest from the fact that it was found that under certain conditions of temperature, pressure, and substance in contact with carbon, hard particles resembling a diamond were produced, which satisfied all the tests for diamond, so far as they could be applied to particles under 1/500 inch in length.

At the time of reading the paper a few tests only had been applied to ascertain whether the particles found were veritable diamonds. Shortly after, however, they were examined by Professor Crookes with electrical discharge in high vacua, and appeared to him to behave in a similar manner to diamond powder.

Tests of specific gravity by immersion of the particles in borotungstate of cadmium and iodide of methylene gave a density of 3-3 to 3-5.

The particles appeared to consist of two kinds—one, irregular opaque black particles; the other, translucent plates resembling flakes of mica, generally of square or irregular shape.
When placed in a cell in iodide of methylene and projected by an electric lantern on the screen, they were clearly seen—the plates appeared to be about 1/500 inch in length and of extreme thinness.

On subjecting the powder to the blowpipe all hard particles disappeared, leaving a yellowish-grey residue, but it should be stated that the powder was not previously levigated to remove the lighter portions, which would account for the residue. 

Read before the Royal Society, June 21th, 1907

Following the subject of my paper of 1888 to this Society, which will be referred to in a subsequent communication, attempts have recently been made to melt carbon by electrical resistance heating under pressure, and the following is a short summary of the results of about one hundred experiments.

The procedure has been on two lines. In the first, carbon is treated in bulk in a thick tube of 8 inches internal diameter of gun steel closed below by a massive pole of steel insulated from but gas tight with the mould and above by a closely fitting steel ram packed by copper rings embedded in grooves in the ram or by leather and steel cups according to whether solids, liquids or gases are to be contained. The bore of the mould is generally lined with asbestos and after being charged the whole is placed under a 2000-ton press, the head and baseplate being insulated and connected to the terminals of a 300-kilowatt storage battery with coupling arrangements for 4, 8, 16 or 48 volts.

It was hoped that the greater thermal and electrical conductivity of steel as compared with carbon or graphite at moderate temperatures would with the help of water jackets keep the outer layers comparatively cool, and that the increased conductivity of the central portions consequent on their higher temperature and conversion to graphite would so centralise the current on the core lying between the poles as to melt it.

Further concentration of current was obtained in the initial stages of heating by packing the central portion with carbon rods on end or by a compressed graphite core, and filling in around with coarsely broken arc-fight carbon, or with wood charcoal (which is a bad conductor until highly heated).

With pressures of about 30 tons per square inch, and currents commencing at 6000 amperes, increasing up to 50,000 amperes, with about 2 volts between the terminals of the mould, the carbon rods were partially converted to graphite and firmly welded together; in the case of the graphite core the flakes were much increased in size.

The heating was in all cases limited by the melting of the steel poles and resulted in short circuits in the mould from the permeation of the asbestos by the molten iron. Neither the internal water jacketing of the poles nor the substitution of copper poles for steel have remedied this trouble.

It appears that the thermal conductivity of the carbon or graphite at or near the temperature of vaporisation is very greatly in excess of that anticipated, or that the rapid transfer of heat is caused by carbon vapour, which appears to have a great power of penetration through carbon at high temperatures. The melting of the poles and the destruction caused by short circuits which reached 80,000 amperes in the mould were not only costly to remedy, but caused contamination of the carbon from the metal of the poles and the insulating material.

In several experiments a nucleus of very soft graphite about 2-1/2 inches in diameter was found in the centre. And in several experiments small masses of iron, highly charged with graphite, were found in varying posi­tions among the carbon or graphite.

This method, however, would probably be more successful if carried out on a much larger scale, as for a given central temperature the transfer of heat to the poles and mould would be less, and water jackets would then prove more effective. It is, however, difficult to construct water jackets to withstand more than 30 tons per square inch, and unless made of hard steel they crush in. The maximum power of the press is 2500 tons, and with the apparatus at hand if the size of the mould was much increased the pressure in the mould would have to be decreased.

Another plan was then adopted of interposing an insulating barrier of some refractory material with a hole in it between the poles, the charge in the first instance being graphite. It was hoped that by means of electrical currents of higher potential and large volume the energy would be so concentrated on the small volume in the neck as to melt it before it had time to form carbides with the material of the barrier.

This was to some extent achieved in that the graphite in the centre was converted to a softer and more flaky nature.

In one of these experiments the barrier was formed out of a block of fused magnesium oxide, specific gravity 3-65, and the pressure in the mould, which was 4 inches in internal diameter, was in this case raised to 100 tons per square inch. The strongest steel poles were required for this pressure, also the mould of gun steel became permanently strained and required reboring after each experiment.

A current at about 12 volts at the terminals in the mould, developing about 100 kilowatts, was turned on for 7 seconds.

The initial diameter of the hole in the barrier was 5/8 inch and the thick­ness about 3/4 inch ( The heat units delivered on to the neck being about four times that required to raise the graphite column through 5000° C, taking the specific heat at 0-5). This barrier was converted to magnesium carbide of a green colour to a radial depth of about 3/8 inch. Thus this magnesium oxide when heated under pressure with graphite readily forms a carbide. The graphite in the centre was altered to large and very soft flakes. Neither the graphite nor the magnesium carbide contained any hard crystalline carbon.

Similar experiments were tried with carbon rods surrounded by silica, and as a guide to the temperature reached, current was turned on of just sufficient voltage to convert the rod to graphite; the mould was then set up afresh and double the voltage applied, when the rod was vaporised and disseminated throughout the molten silica, principally in the form of graphite of very small grain, very little silicon and still less silicide of carbon being formed.

Another series of experiments has been made to investigate the be­haviour of vaporised carbon under fluid or gaseous pressures of about 30 tons per square inch. The general arrangement of the mould consisted of a central carbon rod with a lining of marble; in some cases the space between the rod and marble was packed with coarsely powdered charcoal.

Several compounds of carbon were treated, perhaps the most interesting being carbon dioxide. The liquid was run into the mould and a pressure of 30 tons per square inch applied. It was found that its volume diminished to about 80 per cent., due to its compressibility. Current was then passed through the rod, and the liquid must then have existed as gaseous carbon monoxide in the hotter zones.

When cooled, the liquid and gas were allowed to escape; a sample of this gas on analysis was found to contain 95 per cent, of carbon monoxide and 3 per cent, carbon dioxide, the residue consisting apparently of nitrogen.

As the pressure of 30 tons was maintained throughout the experiment, it would seem that the compressibility of carbon monoxide diminishes rapidly at such high pressures, but this experiment will be repeated and will form the subject of a subsequent paper on the compressibility of liquids and gases. Part of the central carbon was converted to graphite, and in one place there was found a nest of woolly deposited carbon, showing that under a pressure of 30 tons per square inch carbon vaporised in carbon monoxide is deposited in the form of amorphous carbon.


From these experiments several hundred samples have been carefully analysed. In none of the experiments designed to melt or vaporise carbon under pressure has the residue contained more than a suspicion of black or transparent diamond.

In no experiment we have made has there been any sign of the carbon becoming a non-conductor, and the impression derived is undoubtedly that soft crystals of graphite are the resulting stable form of carbon after heating to very high temperatures.

At very high temperatures and pressures graphite has a great tendency to permeate or diffuse into its cooler surroundings. It should, however, be noted that in all the experiments so far made it has been found impossible to exclude from the graphite other substances in the liquid or gaseous state.

Though in many of the foregoing experiments the molten steel of the poles became highly charged with graphite, further experiments have been made to ascertain the influence of pressure upon iron highly charged with carbon. Cores formed of iron rods, iron tubes filled with carbon or with various proportions of iron filings and lamp black, surrounded with various substances such as charcoal, magnesia, olivine, etc., were melted or vaporised and disseminated throughout the charge.

Thus iron highly charged with carbon under a pressure of 30 to 50 tons was cooled at various rates according to its proximity to the sides of the mould, the analysis showing in most cases no residue at all, but occasionally a suspicion of very minute diamond. As a further experiment, a small carbon crucible containing iron highly charged with carbon from the electric furnace was quickly transferred to a steel die and subjected, while still far above the melting point, to a pressure of 75 tons per square inch.

The analysis showed scarcely any crystalline residue and probably less than if the crucible had been cooled in water at atmospheric pressure, and as it would seem that 75 tons or even 30 tons per square inch must be a greater pressure than can be produced in the interior of a spheroidal mass of cast iron when suddenly cooled, the inference from these experiments seems to be that mechanical pressure is not the cause of the production of diamond in rapidly cooled iron.

We hope to be able to communicate further experiments on this subject during the course of next session. I would wish to add that most of the analyses have been made by Mr J. Trevor Cart.

Please help beat cancer - DONATE click above

Unrelated to this post, below is an example of
eclectic science esoterica 

Autunite (hydrated calcium uranyl phosphate) Ca(UO2)2(PO4)2·10-12H2O
Hagendorf South Pegmatite (Cornelia Mine; Hagendorf South Open Cut), Hagendorf, Waidhaus, Vohenstrauß, Oberpfälzer Wald, Upper Palatinate, Bavaria, Germany
Picture width 2 mm. Collection and photo Christian Rewitzer

Main Menu - click above
WARNING - Many subjects outlined within this site are extremely dangerous and are provided here for information only. Please don`t experiment with high voltages or chemicals unless you are fully conversant with safe laboratory practices. No liability will be accepted for death, injury or damage arising from experimentation using any information or materials supplied.