WO1983004016A1

WO1983004016A1 - Process for manufacturing diamond

Info

Publication number: WO1983004016A1
Application number: PCT/US1983/000398
Authority: WO
Inventors: Mark K. Mitchell
Original assignee: Mitchell Mark K
Priority date: 1982-05-20
Filing date: 1983-03-14
Publication date: 1983-11-24
Also published as: EP0110897A1

Abstract

A process for synthesizing diamond from graphite or amorphous carbon which may be ultrapure or contain any desired dopant concentration. Graphite or amorphous carbon is maintained at pressure and temperature conditions where diamond is thermodynamically stable, then the 2sp2 bonds of the graphite or amorphous carbon are ionized by pulses of light of wavelength 250.274 plus or minus 0.154 nm or less. Each appropriately timed cycle of successive ionization and relaxation continuously extends the single diamond crystal, so that, by repetition, diamond of arbitrary size and shape can be manufactured.

Description

PROCESS FOR MANUFACTURING DIAMOND

This invention is concerned. with the production of synthetic diamond.

In the past a number of methods have been found sucess ful in producing synthetic diamond. Perhaps the most widely used is the metal catalyst process developed by General Electric in the 1950's. This process is the subject of U.S. patents, numbers 2,947,608 through 2,947,611, and although many subsequent variations have been developed, all employ the same basic techniques. Graphite or amorphous carbon is placed in the chamber of a belt pressure apparatus with induction heating capability. Also placed in the chamber in various configurations, is a metal which will act as a metal catalyst. Many different metals have been used: iron, nickel, cobalt, ruthenium, rhodium, osmium, irridium, platinum, chromium, manganese, tantalum, or an alloy containing some combination of these metals. The carbon-metal assembly is then subjected to heating to raise the temperature greater than the carbon-metal eutectic. Effectively, the carbon is dissolved into the molten metal (except in the case of tantalum, which does not need to be molten). Pressure is increased until pressure and temperature conditions are above the Berman-Simon line on a pressure vs. temperature graph of the phase diagram for carbon. Typically operating temperatures range from 1200°C. to 2500°C. at pressures of 45 kbar or higher.

Another process which has been reported is the use of explosions to generate the required high pressure and temperature conditions for very short periods of time to spontaneously convert graphite to diamond. Only microcrystals detectable by x-ray diffraction have ever been produced by this method. Many different means of growing additional, diamond onto seed crystals have been proposed. One method employs an ion accelerator to implant carbon ions into an already existing lattice (covered in U.S. patent number 4,191,735). As long as graphitization temperature is not reached, due to the ions' kinetic energy imparting heat to the diamond, the crystal grows from inside.

Seed crystal diamonds have also been extented or grown onto by condensing carbonaceous gas onto the surface of the seed crystal and subsequently subjecting the diamond to temperature and pressure conditions where diamond is thermodynamically stable. The deposited layer undergoes reaction and the carbon in the layer becomes diamond, extending the seed crystal's lattice structure.Repeating the cycle allows the production of larger diamonds than were started with originally, however the growth rates are very low.

Lastly, it is worth noting that lasers have been used to remove flaws in diamond crystals, again avoiding imparting enough energy to the diamond to raise it above graphitization temperature, in which case the diamond will burn.

Diamond crystals produced by the metal catalyst method are limited in size, the largest reported being less than a carat, and usually have atoms of the metal catalyst used incorporated as impurities in the diamond crystal lattice. While diamond produced this way has improved properties for use as grinding media, such uncontrolled impurities are not desirable in gemstones or diamond manufactured for optoelectronic semiconductor, or laser applications. Certainly it is preferable to be able to choose what specific kind of dopant material is to be present and in what concentration and distribution.

By means of this process, diamond crystals, having a single, continuous crystal structure, of any desired size, shape, or doping characteristics can be manufactured. Effective operating pressure can be as low as 18 to 20 kbar, at temperatures of 0°C. to 100° C .

The first step is to maintain graphite or amorphous carbon at diamond stable thermodynamic conditions, that is, any of the pressure-temperature points above the Borman-Simon line on the phase diagram for carbon (see figure 1). Next, the graphite carbon-carbon bonds are broken by ionization from pulses of light, within a small volume of carbon. Then, in the duration between pulses, carbon-carbon bonds are allowed to reform. Since the carbon is being maintained at diamond stable conditions, diamond forms. Sequentially, a light pulse ionizes the graphite carbon-carbon bonds and the carbon-carbon bonds of diamond reform in the time between pulses. Repeating this ionization-relaxation cycle in adjacent, unconverted carbon by subsequent light pulses, uniformly spaced apart in time, permits sucessive, continuous extension of the diamond crystal so formed.

Figure 1 is a pressure vs temperature graph showing the crystal phases of carbon. All the. pressure-temperature points above the Berman-Simon line (labeled "b") in the region denoted by the letter "a" present conditions where diamond is thermodynamically stable. Conversely, the pressure-temperature points in the region denoted by the letter "c" represent the thermodynamic conditions where graphite is stable. The pressure scale on the graph is in kilobars, abbreviated kbar, and the temperature scale, as indicated, is in degrees centigrade. Whenever the phrase "conditions where diamond is thermodynamically stable" is used, it refers to any pressure-temperature point in region "a" .

It should be further understood that any carbon-carbon bonds formed under graphite stable conditions will have the same characteristic type of hybrid, electron bonding orbital in the bonding carbon atoms, called 2sp². In the same fashion any carbon-carbon bonds formed under diamond stable conditions will also only have one characteristic type of hybrid electron orbital called 2sp³. Therefore, the carbon-carbon bonds of graphite (again true for all carbon existing in that crystal phase, and therefore true for amorphous carbon as well) will hereafter be simply designated 2sp bonds; and the carbon-carbon bonds of diamond will simply be designated as 2sp³ bonds. Specifically what the terminology means, is that in the case of graphite, one of each carbon atom's 2s electron orbitals and two of its three 2p orbitals have hybridized to form three equal energy bonding orbitals, each a type 2sp². In diamond, one of each carbon atom's 2s electron orbitals and all three of its 2p orbitals have hybridized to form four equal energy bonding orbitals, each a type 2sp³.

Each of the 2sp ² carbon-carbon bonds in graphite have the the same energy, that is, approximately 4.954 ± 0.003 electron volts. Whereas diamond having at least two forms, duodecahedral and octahedral, has 2sp³ carbon-carbon bonds which have approximately 7.23 ± 0.38 eV (electron volts). energy per orbital or bond in duodecahedral diamond, and approximately 8.82 ± 0.32 eV energy per orbital or bond in octahedral diamond. Both duodecahedral and octahedral diamond have type 2sp³ carbon-carbon bonding orbitals (all energies are ± 0.5 eV). In this process, photons of light ionize the electrons in the bonding orbitals by means of the photoelectric effect. This simply requires that each electron so ionized must be knocked out of its bound state by a single photon having an energy equal to or greater than the amount of energy by which the electron is bound. Since the energy of a photon is given by E=(hc)/λ (where E is the photon's energy, h is Planck's constant, c is the speed of light, and λ is the wavelength of light), light with a wavelength of 250.274 ± 0.154 nm (nanometers) or shorter has photons energetic enough to break the carbon-carbon bond of graphite. Similiarly, light of wavelength of 165.736 ± 5.736 nm or shorter will break the 2sp³ bonds of duodecahedral diamond; and, light of wavelength 134.921 ±

5.699 nm or shorter will break the 2sp³ bonds of octahedral diamond.

Light of wavelength 250.274 ± 0.154 nm is most effective at breaking the 2sp² bonds of graphite by ionizing the electrons in the bonding orbitals; since, at this wavelength, each photon has precisely the same amont of energy as that by which the electron is bound. As a result, the best light source is a mercury vapor laser which has an output of coherent, monochromatic, pulsed light, whose peak wavelength output is 250.274 ± 0.154 nm. Although, any light source, which puts out light of wavelength of 250.274 ± 0.154 nm or shorter can be used.

The reason there must be a pulse cycle to the light incident on the starting material, is to not only provide for the the breakage of the 2sp² graphite carbon-carbon bonds, but also to allow time for the ionized electrons to dissipate their kinetic energy as heat to the surrounding caxbon atoms and be recaptured by the ionized atoms, thus forming new 2sp ³ carbon-carbon bonds. Both ionization and relaxation segments of the process cycle are necessary.

Any intensity light pulse can be used, but the less intense the pulse the fewer bonds will be broken and less graphite will be converted to diamond per pulse cycle. Slow accumulative growth will result. Also, the pulse can be of any duration; however, after the graphite already exposed to the beam of light is already ionized, free electrons will preferentially absorb the rest of the incoming light, much as a charged gas would. This only serves to raise the temperature of the starting material, making for a longer relaxation time. Short, intense pulses are optimal, of duration of the order of microseconds or less. Relaxation times will always be longer than the duration of the light pulse, generally by a period of time at least an order of magnitude longer. To optimize, that is, to shorten as much as possible, the necessary relaxation time, the carbon must be adequately cooled.

Each subsequent pulse cycle converts graphite or amorphous carbon adjacent and contiguous with the diamond previously formed. Since the newly converted graphite adopts the lattice structure of the diamond already present, that diamond crystal is extended or grown onto in a continuous fashion. As such, using repeated iterations of the process cycle, a diamond of any size and shape with a single diamond crystal structure is produced. A diamond seed .crystal can also be added onto, continuously extending the diamond lattice of the seed crystal, by this same method.

It should be emphasized, however, that the active part of the pulse must be of duration long enough to provide ionization sufficient to break all the 2sp² graphite bonds of all the carbon atoms within the interaction volume, so thata uniform,homogeneously excited plasma results. Therefore, further excitation is necessary to insure that all carbon atoms whose chemical bonds are broken, are saturated (i.e. have all of their bonds broken) and are in the same average excited set of states before relaxation begins. Cooling provides a condensation gradient (particularly since diamond conducts heat so well, and so will be cooler than the insulating graphite or amorphous carbon powder, which does not conduct heat well) so that once a nucleation site for crystallization forms, or such sites are already present in the form of available open lattice sites of diamond present from a previous conversion cycle or an initial seed crystal, crystallization will proceed from the(those) favored site(s) throughout the remaining interaction volume uniformly. Thus, the formation of a single crystal structure is assured by uniform excitation, uniform pressure and the resulting crystallization propagation along a condensation gradient.

Any number of schemes can be employed to permit access of the subsequent light pulses to the successively adjacent, contiguous volumes of graphite (e.g. projection of subsequent pulses through other radiation ports at different angles). Yet, the best arrangement is to use light of wavelength of 250.274 ± 0.154 nm or shorter, but greater than 165.736 ± 5.736 nm, so that subsequent pulses can be projected through the previously formed diamond and still not be energetic enough to break the 2sp ³ carbon-carbon bonds of the diamond.

This insures that the adjacent volume of graphite is indeed contiguous with the diamond previously formed, which is neces sary if a single crystal diamond is to be produced. So while it is possible to use light of wavelength of 165.736 ± 5.736 nm or shorter to break the 2sp ² graphite bonds, unless a projection scheme is found so that subsequent pulses are not in- cident on the diamond already present, the diamond 2sp³ bonds will be effected also. In the case where polycrystalline dia mondis converted into a single crystal diamond, light of wavelength 165.736 ± 5.736 nm or shorter for duodecahedral diamond or light of wavelength 134.921 ± 5.699 nm or shorter for octahedral polycrystalline diamond must be used. However, the light pulses cannot be projected through crystal already treated. Otherwise, except for the starting material and the shorter wavelength restriction on the light which is used, the process is the same. The graphite or amorphous carbon starting material can be ultrapure, inwhich case the diamond crystal will have no impurities incorporated in it. Certainly no impurities are introduced by the conversion process. Any desired dopant material (e.g. boron, phosphorous or aluminium) can be intermixed, in any arbitrary concentration and distribution, as a simple mixture or chemically bonded to the carbon starting material in a specific fashion, before being subjected to the conversion process. The resultant diamond produced will then have the prearranged concentration or distribution of dopant material incorporated into the crystal. Although it is the process itself and not the apparatus which is the subject of this patent, the following is a description of a specific apparatus for performing this process.

A belt style press with two conical punches which punch through an annular ring, thus forming a chamber which is packed with pyrophillite (e.g. talc), is used to subject graphite in the chamber to a pressure of 20 kbar. The ring has a cooling jacket and one or more radiation windows, transparent to the wavelength of light being used (e.g. polycarbonate or quartz). Graphite in the chamber is ionized by pulses or light from a mercury vapor laser whose output is predominately light of wavelength 250.274 ± 0.154 nm. Pulses are 0.7 microseconds in duration separated by 2 milliseconds. The ionization-relaxation cycle is initiated and diamond forms. Successive pulses are projected through previously formed diamond to convert graphite contiguous with the diamond. The diamond is augmented by each pulse cycle's action, until a single crystal diamond is produced. A variation is to place a diamond seed crystal against the inside face of a radiation window and pack graphite or amorphous carbon around it. Then the pulses of light are projected through the diamond, the result being the same.

Throughout these operations, temperature is controlled to be from 0°C to 100°C by the use of water in the cooling jacket.

Optimization of the conversion process can be achieved by starting with graphite powder of the smallest possible average crystal size (i.e. amorphous carbon), to delimit as much as possible the quantum mechanical degrees of freedom of each part of the initial substrate, and loading the starting material under hard vacuum conditions as may be obtained using a turbo-molecular or cryogenic vacuum pump, in order to exclude the incorporation of optically absorptive N₂ (nitrogen gas) from air into the prospective crystal. The prime operational wavelengths for the molecular photoionization process for manufacturing diamond herein described are from 250 to 225 nm for resons stated above.

Lastly there is no reason why this.kind of process, i.e. selective molecular photoionization of the chemical bonds of dense materials under specific pressure-temperature conditions, could not be applied to other elements, to manufacture novel materials with potentially useful properties.

Claims

I make the following claims:

1. That diamond having a continuous, single crystal structure of any arbitrary size, shape and doping characteristics is manufactured by a. subjecting graphite or amorphous carbon, ultrapure or intermixed with any desired dopant material of any concentration and distribution, to pressure and temperature conditions where diamond is thermo dynamically stable, maintaining these conditions

(unifrom in pressure and constant in temperature by cooling through subsequent process steps), and b. ionizing the 2sp² bonds of the graphite or amorphous carbon with pulses of light of wavelength 250.274 ±

0.154 nm or less, the duration of each pulse adequate to provide complete, homogeneous ionization in the interaction volume and the time between pulses suf ficient to allow relaxation and forming of the 2sp³ bonds of diamond, whereby c. inducing the same process of ionization and relaxation in the graphite or amorphous carbon adjacent to and contiguous with the diamond previously formed serves to continuously extend the diamond crystal, and that, d. repetition of this process cycle any number of iterations continuously extends the single diamond crystal so that a single diamond crystal of any desired shape and size is produced.

2. That any single crystal diamond can be extended and augmented by said process.

3. That any polycrystalline diamond can be converted to single crystal structure by maintaining it at pressure and temperature conditions where diamond is thermodynamically stable and using said process with pulses of light of wavelength 165.736 ± 5.736 nm or less for duodecahedral polycrystalline diamond and of wavelength 134.921 ± 5.699 nm or less for octahedral polycrystalline diamond.