ß-silicon nitride, Si3N4, is an exceedingly hard hexagonal phase ceramic that has finally been obtained as its cubic polymorph, Nature 400 312;340 (1999). Theoretically stiffer ß-carbon nitride, C3N4, may be harder than diamond, though Solid State Commun. makes a good case for C4N3 in preference, replacing a non-bonding N-N lone pair juxtaposition with a C-C bond.
J. Am. Chem. Soc. 121 4696 (1999)
Chem. Mater. 8 318 1996
Science 271(5245) 53 (1996)
Science 261 334 (1993)
Science 245 841 (1989)
Phys. Rev. B 41 10727 (1990)
Solid State Commun. 96 321 (1995)
Phys. Rev. B 32 7988 (1985)
Mater. Chem. Phys. 47 109 (1997)
Mater. Chem. Phys. 43 1 (1996)
Adv. Mater. 9 877 (1997)
(Both carbon nitrides are puckered layers of fused 8-membered rings bonded by single bond pillars. Neither one looks anywhere near as atomically dense as diamond. Organikers will be curious as to how trigonal nitrogen and tetraheral carbon work out their valence electrons.)
Graphite and elemental nitrogen are deep thermodynamic holes, as is diamond under pressure. Alternating carbon-nitrogen 3-D lattice synthesis is assigned to engineers (thing makers) not scientists (stuff makers). More studies are needed.
Federal funding supports the tiniest latent advance and smallest publishable byte, allaying risk and assuring success (bantam or imaginary). Charles Lieber (http://magic.harvard.edu/) shook the money tree. US Pat. 5,840,435 "Covalent carbon nitride material comprising C.sub.2 N and formation method" (re his 1993 Science paper; RF-excited atomic nitrogen in helium kisses laser-ablated carbon) cites Air Force Office of Scientific Research grants.
Organic chemistry offers precursors with high N/C ratio, diamond stability field 850,000 psi/1450 C, catalysts, nitride solvents, and release for excess atoms. Organikers bamboozle equilibrium thermodynamics and hoodwink reaction kinetics. Good precursors possess alternating C-N bonds. No problem.
Lead cyanamide [Pb(2+)] [(-)N=C=N(-)] is a wild and hairy substrate. Will heat and pressure move electrons, extrude lead (density increase to 11 g/cm^3, movement lubed by soft metal), and birth ß-C3N4? Excess nitrogen exits as gas, or bring things into stoichiometry via intimate admixture with Pb(CN)2. /_\(PV) is energy. Give it a big negative /_\(PV) to drive Miss Daisy home.
Cyanamide, H2NCN, is cheap. Guanidine, (H2N)2C=NH, is stable as protonated salts. 3:1 N/C is in there! Try it as the cyanide, C2H6N4. Guanidine cyanimide is C2H7N5. If a dibasic salt forms it is C3H12N10. Guanidine carbonate is commercial, C3H12N6O2.
Melamine, C3H6N6, is common. The analog from cyanuric chloride and hydrazine is C3H9N9, but R-NH-NH2 may extrude molecular nitrogen under forcing conditions and will grab CO2 from air.
React melamine with three moles of cyanuric chloride giving a first generation dendrimer. Then six moles of melamine for second generation, 12 moles of cyanuric chloride for third, 24 moles of melamine for fourth... to grow molecular torus -N=C-N=C-N-C=N-C- with steric hindrance deterring its aromatic rings from coplanarizing and graphitizing. The slowly growing dendrimer has limiting elemental composition C3N4.5H1.5 (or 2(C3N4) + NH3; not so bad). If you really wail on the synthesis to get fully arylated peripheral nitrogens, reacting each of melamine's pendant -NH2 groups with two moles of cyanuric chloride, then the rapidly densifying dendrimer has limiting elemental composition C3N4. Ta da! No muss, no fuss, no waste. Just reshuffle bonds.
For more nitrogen and elbow room, react one cyanuric chloride with three moles of hydrazine giving a triply amino-extended melamine analog core. Then three cyanuric chlorides for first generation, six analogs for second generation, 12 cyanuric chlorides for third generation, 24 analogs for fourth... to grow a less crowded torus -N=C-N=C-N-N-C=N-C- again with steric hindrance deterring graphitization. Longer arms allow another generation or two before crowding overcomes reactivity.
A dendrimer has empty space toward its center and an atomically dense surface. Give it a squeeze and the hyperbranched polymer molecule resists, them fails and implodes. You don't drive a nail by pushing on it. You strike it.
Extraneous atom disposal: Excess nitrogen addresses literature complaints of getting enough combined vs graphite formation. Hydrogenated substrates compel aliphatic structure versus graphite. Hydrogen diffuses away on its own, or is tickled with metallic titanium (carbide, nitride, and hydride phases) or palladium (a sponge for hydrogen).
Expediting kinetics: Graphite into diamond needs a carbon solvent/catalyst for useful kinetics at achievable pressures and temperatures. Li3N is a curious solvent for ß-carbon nitride synthesis. Dispersed titanium nitride, carbide, and hydride interstitial phases may provide nucleation centers for synthesis and crystallization as iron or cobalt highly dispersed catalysts grow carbon nanotubes.
Making ß-C3N4 is like making diamond. H. Tracy Hall grew diamonds at General Electric in 1956. His kids are in the business: H. Tracy Hall, Jr., http://www.novatekonline.com/
US Pat. 4124690, 4174380, 4184079 teach synthetic Type Ib diamond soaked two days in the diamond stability field allows nitrogen diffusion from atomic dispersion to platey aggregates as in Type Ia natural diamond. Lattice defects from energetic particle irradiation help. Man-created, treated, natural - same stuff.
Does HP/HT ß-C3N4 disproportionate to diamond and nitrogen through lattice migration? Make it on the fly by fast quench staged implosion (abrasive dust only and no sintered forms). Graphite becomes diamond dust in 40% yield this way. (Copper jacket - ferrous alloys encourage graphitization). Lattice annealing eats you. Will boron effect lattice defect pinning?
Give your precursors a big hug. If you are quick about it, get ß-C3N4. Idiot engineers and their interminable incremental optimizations will be astonished.