Carbon atoms form very strong bonds to each other, yielding materials like: (i) crystalline graphite, diamond and their many “amorphous” hybrids; (ii) crystalline forms of giant closed–surface molecules such as the fullerenes; and (iii) liquid and gas phases which have molecular contents which are complicated and not yet defined or understood. Because of the high bonding energy the melting and vaporization temperatures of the solid forms are very high, and the activation energies required to transform one solid form to another are large. One consequence is that at lower temperatures the different solid phases may continue to exist metastably far into a P, T region in which another solid phase is the thermodynamically stable one.

In the thermodynamic sense the vapor pressure line of graphite, the graphite/liquid/vapor triple point, the graphite melting line, the graphite/diamond equilibrium line, and the graphite/diamond/liquid triple point are quite well established. Data for the melting temperature of diamond vs. pressure are sparse and rough, but they indicate that the melting temperature increases with pressure,-in agreement with some theories. Although carbon should transform to a solid metallic state at very high pressures, experimental evidence shows diamond to be stable to over 400GPa, and theoretical calculations indicate that it could be the stable form up to pressures of 1200 to 2300GPa. Attention is given to the solid state transformations which can take place when graphite is compressed and heated along different P, T paths under different conditions.

Shock wave synthesis of diamond was an unexpected result of experiments designed to explore the effects of shock waves on a variety of materials. The initial announcement in 1959 was controversial; shock synthesis of diamond had been shown to be unlikely, on the basis of kinetic arguments. Jamieson confirmed the identification and suggested a diffusionless mechanism, c-axis compression of rhombohedral graphite. Subsequent work has provided strong evidence that shock wave synthesis of cubic diamond is a conventional thermally activated nucleation and growth process. Thermal inhomogeneities provide the requisite high temperatures; quenching via thermal equilibration is implicit in the process. Shock synthesis of adamantine BN phases appears to be quasi-martensitic; a martensitic mechanism may partially account for the Lonsdaleite (hexagonal diamond) observed in some meteorites and in some artificial shock products. Diamond is also formed as a detonation product in oxygen-deficient explosives. The polycrystalline product of shock synthesis is similar to natural carbonado. The association of carbonado with an ancient giant impact crater is noted.

Starting from properties of the carbon atom, it is possible to explore the structural stability and properties of various forms of carbon. A discussion of the theoretical approach and the relevant atomic characterisitcs will be given followed by example of applications to existing and hypothetical materials formed from carbon and related elements.

Diamond quality is strongly coupled to growth rate. Incorporation of nondiamond (sp2) carbon and morphological instabilities both increase with increasing growth rates. The intersection of twins with the growth surface produces re-entrant corners that enhance growth in the plane of the twin. Morphology and the development of texture both depend on substrate temperature and methane concentration and hence on growth rate. Experimental evidence and modeling results that relate growth rates and quality to controllable process parameters are reviewed.

A brief review is given of recently published work on the fracture and plastic flow of natural and synthetic diamond crystals when deformed by softer impressors at elevated temperatures. Under these conditions, a brittle-ductile transition (BDT) temperature for the different types of diamond has been established.

In this paper we show that below the BDT temperature, brittle ‘chatter’ cracking on {111} cleavage planes accounts for anisotropic wear provided that the mean contact pressure exceeds about 10 GPa. Above the BDT temperature, titanium diboride sliders develop contact pressures sufficient to cause extensive plastic deformation of the diamond specimens but insufficient to produce cleavage fracture. However, repeated sliding with TiB2 does lead to cumulative plastic flow preceding fatigue type fracture on {110} planes. Using cubic boron nitride and diamond sliders above the BD T temperature leads to a combination of both plastic deformation and {110} cracking. The coefficient offriction is a maximum in <100> and a minimum in <110> directions. and the measured scratch hardness, which confirms that the <100> directions are softer than <110>, is shown to be consistent with the predictions of a resolved shear stress model which was de velopedffor explaining anisotropy in all crystals.

Attempts were made to produce bands of hexagonal diamond (Lonsdaleite) by high-temperature indentation of cubic diamond. (011) wafers of type I diamond were indented over the temperature range 1000–1300°C and the microstructure of the indentation plastic zone investigated by transmission electron microscopy (TEM). No hexagonal diamond was produced; instead, the plastic zone consisted of arrays of dislocations lying on the {001} planes. A possible mechanism for the generation of such dislocations, and reasons for the absence of hexagonal diamond, are discussed.

High precision force-displacement measurements of various forms of glassy and graphitic carbon have been made with a Berkovich and spherical tipped indenter. The force-displacement data show almost complete recovery despite contact pressures exceeding a considerable fraction of the elastic modulus. With a spherical tipped indenter the transition from elastic to inelastic response could be readily identified. It was found that the extent of the hysteretic response with a spherical indenter increased with increasing load. Only minor differences in the force-displacement response of crystalline graphite indented normal to the basal plane and the glassy carbon was identified. A simple model for interpreting the force-displacement data is developed. These results are compared with indentations made on diamond like carbon films on various substrates.

Two areas of thin film property measurement are addressed. The first is that of flat films, either on a substrate or free-standing. The film properties only are of interest. Therefore, when the film remains attached to a substrate during testing, an appropriate analysis is used to subtract the effect of the substrate. The films under test are prospective protective coatings and ‘window’ materials for infrared applications, namely CVD diamond (Hot filament Assisted, HFACVD and Microwave plasma assisted, MPACVD) and Germanium carbide (Ge:C). The mechanical properties under investigation are the Young's modulus and the internal film stress.

In the second case the substrates are small diameter fibres and wires coated with CVD diamond. The mechanical properties measured were composite, containing contributions from both the substrate and the film. These coated fibres and wires, have possible applications as reinforcement phases in the production of composites. They are silicon carbide (SiC) and Tungsten (W) of diameters varying between 10 and 125μm. A technique has been developed to measure the Young's modulus of individual coated fibres.

Control of the mechanical properties of CVD diamond is essential to achieve optimal performance in various applications. While several methods have been applied to the measurement of the Young's modulus of thick-film CVD diamond, in general these methods are not suitable for diamond characterization on a production scale. In addition, many of these methods cannot determine the shear modulus (or Poisson's ratio), which is necessary for a complete description of the elastic properties. We have developed a simple dynamic resonance method for determining both the Young's and shear modulus of free-standing CVD diamond in the shape of rectangular plates or round disks. The specimen is supported along nodal lines of flexural or torsional modes. Oscillations induced by impact from a falling ceramic bead are sensed by a microphone, and the resonant frequencies are determined by a signal analyzer. The Young's and shear modulus are calculated from the frequencies of the fundamental flexural and torsional modes, respectively, using quasi-analytic formulas. CVD diamond grown by several methods routinely achieves Young's and shear modulus values above 1000 GPa and 500 GPa, respectively, in good accord with theoretical values for pure polycrystalline diamond.

The density and elastic properties of a 1.9 μm thick polycrystalline diamond film deposited on a silicon substrate were measured by surface acoustic wave (SAW) spectroscopy. A density of 3.45±0.05 g/cm3, Young's modulus of 940 ± 20 GPa and Poisson's ratio below 0.12 were determined from the dispersion of a broadband coherent surface acoustic wave pulse propagating in the layered system. The surface wave pulses were generated using a ns UV laser pulse and detected with a piezoelectric foil transducer.

We have grown amorphous hydrogenated boron carbide thin films by rf plasma decomposition of diborane and methane. The chemical composition, infrared absorption, optical absorption, and microhardness of these thin films were measured. As a function of increasing diborane concentration in the feedstock, we observe increasing boron and hydrogen concentrations, increasing infrared absorption at 1330 cm−1 due to boron icosahedra, increasing optical bandgaps, and an unchanging microhardness in the grown films. The microhardness should have decreased due to the increasing hydrogen concentration; this expected decrease may have been balanced by an increased microhardness due to the boron icosahedra.

Measurement of stress in CVD diamond by Raman spectroscopy and by the substrate curvature method is discussed. A correction to the commonly applied Stoney thin-film equation for the substrate curvature technique is presented for coatings with a large stiffness mismatch with the substrate. Stress measurement by Raman spectroscopy is complicated by factors such as temperature, domain size, non-hydrostatic stress, and degeneracy lifting. With proper consideration of these complicating factors, substrate curvature and Raman spectroscopy stress measurement results can be reconciled with the predictions based on thermal modeling.

Both diamond and amorphous carbon films can contain substantial residual stresses (up to 10 GPa and higher) due to a combination of thermal and intrinsic effects. Residual stresses in the films limit the maximum thickness to which films can be grown and can lead to weakening and/or failure of the film/substrate interface. In addition, a quantitative measurement of the residual stress in the film is required for an accurate determination of the film/substrate interface fracture toughness. Spectroscopic measurements of residual stresses in diamond and amorphous carbon films based on frequency shifts of Raman-active phonon modes are presented. The spatial resolution of the technique is 1 μm in the lateral direction. In transparent materials such as single crystal diamond, stress profiles with 10 μm resolution in the axial direction can be measured. Examples of such stress profiles are presented for natural and CVD-grown single crystals. Raman measurements of thermal stress in thin polycrystalline diamond films on metal substrates are presented. Initial, spatially resolved measurements of large, compressive intrinsic growth stresses in high-hardness amorphous carbon films grown by ion beam techniques are presented. The accuracy of the Raman residual stress measurement depends on the film structure and is ±60 MPa for single crystal diamond, ±120 MPa for polycrystalline diamond, and ±500 MPa for hard amorphous carbon.

MicroRaman spectroscopy has been used to determine residual stress in polycrystalline diamond films grown by CVD process in a DC arc plasma from CH4/H2 gas mixtures. A variety of substrate materials, including silica glass, Si, SiC, Mo, Cu, Ni, Fe-Ni alloy, WC-Co and steel, were selected in order to extend as much as possible the range of the substrate thermal expansion coefficients (CTE). Diamond adhesion to catalytically active materials (Ni, Fe, Co) and Cu was improved with a thin buffer layer of CVD tungsten. The observed Raman peak shifts were converted to stress values σmeas in the framework of biaxial stress model. Compressive stress has been found in diamond films at all substrates, except SiO2, which provided a tensile stress up to +2.3 GPa. The maximum compressive stress has been detected for Ni substrate σmeas =-11.4 GPa at room temperature and σmeas =-14.3 GPa at T=78 K. A linear dependence of stress on temperature was found at Ni in the temperature range 78–520 K, in good agreement with prediction of thermal stress model. The splitting of diamond peak to singlet and doublet becomes observable at σ>8–9 GPa. The measured stress correlates in general with calculated thermal stress, increasing with CTE value of respective substrate materials, copper being the only exclusion because of its high plasticity.

A major problem for diamond coating applications is that diamond films tend to exhibit poor adherence on many. substrates and typically disbond at thicknesses of the order of few micrometers due especially to residual stresses. Residual stresses in diamond are composed of thermal expansion mismatch stresses and intrinsic stresses induced during film growth. Diamond films were deposited in a classical microwave plasma reactor from hydrocarbon-hydrogen-oxygen gas mixtures. Thermal stresses were directly calculated from Hook's law. On silicon substrate, intrinsic stresses were deduced by difference from measurements of total stresses either by the curvature method or by X-ray diffraction using the sin 2ψ method. These investigations allow us to discuss the origin of the intrinsic stresses. The residual stress level was also investigated by Raman spectroscopy as a function of the deposition conditions and substrate materials (SiO2, Si3N4, Si, SiC, WC-Co, Mo and Ti-6A1-4V). We show that the thermal stresses are often preponderant.

Raman spectra in CVD diamond films were measured at room temeprature and at high temperatures up to 1200 K. With micron spatial resolution, variations of Raman line from different crystals in the same sample were observed. Both single-peak and multiple-peak analyses were used to assess the stress distribution within the CVD diamond film. In addition, the evolution of Raman line position, line width, and line intensity was monitored as a function of annealing temperature and isothermal holding time. Both (100) and (111) CVD diamond crystals were studied. The detailed evolution of the Raman line width was found to depend on the crystallographic orientation. The temperature dependence of the Raman spectra for CVD and natural diamonds are compared.

Adhesion measurement of brittle coatings on ductile substrates is examined with emphasis on diamond–coated metals. The experimental aspects of determining film adhesion is discussed through measurement of the interface toughness by a brale indentation method.

The mechanics of thin films are used to define quantitative procedures for predicting interface decohesion motivated by residual stress. The emphasis is on the role of the interface debond energy, especially methods for its accurate and reliable measurement. Experimental results are reviewed and possible mechanisms are discussed.

It is shown that the blister testing technique can be used to measure the adhesion of thin films to their substrates. A brief discussion of blister test mechanics is presented here, leading to a simple equation relating adhesion to the height of the blister and the pressure causing it to grow. Blister test data for plasma-enhanced CVD diamond films on Si substrates have been analyzed using this relation. The tests show adhesion energies of 1.8– 2.6 J/m2.

Poor adhesion of diamond films limits the use of CVD diamond films as coatings for cutting tools. The adhesion of these films is limited by stresses in the film caused by thermal expansion mismatch between the substrate and the film and by voids present at the interface due to the morphology of the crystal growth. A three step process of making diamond composite films has been developed, involving nucleation of individual diamonds on the substrate, electroplating a metal binder in the voids between the crystals, and lastly growing a complete film over the composite layer. The metal binder acts both to fill the voids at the interface and to absorb energy during fracture processes at the interface. Diamond growth was performed in a DC Triple Torch reactor using a mixture of methane and hydrogen with a molybdenum substrate. Measurements to determine the amount of improvement of the film adhesion have been performed. These tests include indentations using conventional hardness testing equipment and four point bend tests with the film in tension and compression. A correlation is shown between the plastic zone of the substrate and the area of the film which delaminated during indentation. Bend tests with the film in tension did not delaminate the film, instead the film underwent intergranular fracture. Bend tests in compression act similarly to pile up around an indentation, and cause film delamination. Residual stress measurements in the single step film show a compressive stress of 650 MPa.