3.9 c. Nanocrystalline diamond synthesized from C₆₀ (N.A. Dubrovinskaia, L.S. Dubrovinsky, F. Langenhorst, S.D. Jacobsen/Washington DC and C. Liebske)

Since the first artificial diamonds were manufactured in the mid 1950’s, various methods (ranging from direct solid-state transformation of graphite under static or shock pressure to chemical-vapour deposition) for diamond and DLC (diamond like carbon) synthesis under variable pressure and temperature conditions have been explored, resulting in materials with properties approaching those of natural diamonds. Ultrafine diamonds with grain sizes of 5-10 nm were synthesised by explosives and may display excellent properties as surface coating for metals. Polycrystalline cubic diamond synthesised by direct conversion of graphite at static high pressures and temperatures is ultrahard. Compression of fullerene C₆₀ under non-hydrostatic pressures to 25-30 GPa at room temperature has resulted in direct transformation to diamond, probably with small crystallite sizes, but the amount of synthesized material was insufficient for detailed characterisation of the structure or mechanical properties and the reproducibility of such experiments has come under scrutiny. Numerous nanocluster-based phases with presumably 3D-polymerized fullerite structures have been synthesized from C₆₀ and carbon single-wall nanotubes at pressures above 13 GPa and temperatures above 800 K. These new materials show impressive mechanical properties and some of them appeared to be harder than diamond. However, subsequent studies did not confirm the exceptional hardness of fullerites.

In course of the systematic studies of direct transformation of low-density carbon materials into diamond, we synthesized nanocrystalline diamond using C₆₀ as a starting material. The new material has a number of unusual and potentially important properties.

We conducted a series of experiments in 1200- and 5000- tonne multianvil presses, using powdered graphite, amorphous carbon, C₆₀, and natural diamond as starting materials. The samples were contained in a Pt capsule to avoid undesirable reactions with carbon. The C₆₀ sample compressed at room temperature to 20 GPa came out as a black non-transparent brittle cylinder. X-ray diffraction and Raman spectroscopy show, in good agreement with previous observations, that fullerene preserves its structure (lattice parameter of the cubic cell reduced from 14.041(3) Å for pristine material to 13.856(3) Å for the product phase) after high-pressure treatment. All carbon phases processed at simultaneous high pressures and temperatures transformed to transparent (in the case of amorphous carbon, graphite, and C₆₀) or semi-transparent (when diamond powder was used) materials. Samples obtained from powdered diamond and amorphous carbon are slightly grey in colour, while materials obtained from graphite and C₆₀ are yellowish. SEM and ATEM studies show that at the conditions of our experiments, platinum did not react with carbon, the samples were not contaminated and contained only pure carbon. X-ray powder diffraction patterns of all recovered samples are dominated by reflections from diamond (Fig. 3.9-3), however, materials synthesized at temperatures below 2600 K from amorphous carbon, graphite, and diamond exhibit several additional small reflections at ~2.19 Å, 1.92 Å, and 1.50 Å (Fig. 3.9-3). These reflections correspond to lonsdaleite (2H diamond polytype) and could result from either a small amount of this phase or disorder of the diamond structure along [111]. X-ray and electron diffraction patterns from the samples synthesized from C₆₀ contain (in addition to the diamond-structure reflections at 2.0497, 1.2552, and 1.0705 Å) diffraction lines at 2.141, 1.9195, 1.6292, 1.3670, 1.1578, and 1.0519 Å (Fig 3.9-3). All these lines can be indexed in the framework of a hexagonal unit cell with lattice parameters a = 2.510(1) Å and c = 12.301(3) Å. Such diffraction data are thus consistent with theoretically predicted 6H diamond-like polytype.

TEM images, as well as estimation from the broadening of X-ray diffraction lines by the Williamson-Hall method, show that bulk samples of diamond obtained at P=20 GPa and T=2300 K from C₆₀ consist of crystallites of 5 to 12 nm. Additional synthesis experiments indicate that nanocrystalline diamond forms independent of the cooling rate. Due to the very small grain size of the nanocrystalline material, high-resolution TEM (HRTEM) images were difficult to achieve, but where possible, HRTEM images show that the nano-crystalline diamond synthesised from C₆₀ has an ideal structure, free of stacking faults or other defects (Fig. 3.9-4). The EELS spectra are also perfectly compatible with the diamond structure (Fig. 3.9-5) and suggest that sp³ hybridization prevails even across the numerous grain boundaries. Hardness measurements of superhard materials like diamond or DLC are problematic, because the measurements are based on the assumption that the tested material plastically deforms. Obviously, the hardness of the indenter should exceed the hardness of the tested material and this requirement limited our ability to measure the hardness of synthesized nano-diamond. It is known, for example, that the hardness of the (111) face of the type IIa diamond (“hardest” diamond face) so far has not been measured, because it is not possible to make an indentation on this face. In our case, a tip of the diamond of a Vickers-type indenter does not make any scratches or indentations on the surfaces of nano-diamonds at loads up to 500 g. Even using SEM we can not see the mark of indentor. So, we conclude that nanodiamond material is at least as hard as usual bulk diamond.

A bulk sample of nanocrystalline cubic diamond with crystallite sizes of 5-12 nm was synthesised from fullerene C₆₀. Its properties were studied using X-ray diffraction, Raman, IR spectroscopy, and HRTEM. It was found that the material has a unique Raman spectrum different from that of usual diamond, and the thermal stability of nano-diamond produced from C₆₀ is at least 300 K higher than that of normal diamond. Such properties of the new material could relate to the small dimensions and sharp size distribution of crystallites (5-12 nm).

Fig. 3.9-3: Examples of diffraction patterns of carbon materials synthesized (a) from graphite at 24 GPa and 2400 K, (b) from natural diamond at 20 GPa and 2300 K, and (c) from C60 at 20 GPa and 2300 K (“D” for diamond reflections, “L” for lonsdaleite (or 2H diamond polytype), “6H” for 6H diamond polytype). Insert shows enlarged plot of pattern c with vertical bars corresponding to the position and intensities of the diffraction lines of 6H polytype of diamond.

Fig. 3.9-4: HRTEM image shows that individual crystallites, which comprised the bulk of the nano-crystalline sample synthesised from C60, have the ideal diamond structure, free of stacking faults or other defects.

Fig. 3.9-5: The EELS spectra from ptestine C60 (bottom); from polycrystalline diamond obtained from graphite at 20 GPa and 2300 K; from the nanocrystalline diamond produced by direct conversion of C60 at the same conditions. Spectrum of the nanodiamond sample (top) treated at 1800 K in forming gas atmosphere does not shoe any sign of sp2- bonded carbon.