Mark Philip D'evelyn

Mechanism of surface smoothing of diamond by a hydrogen plasma

Abstract As-polished diamond (100)- and (111)-oriented single crystals and natural diamond powders of 0.12–25 μm diameter were treated in atomic hydrogen generated by a microwave plasma or by a hot tungsten filament. Post-treatment atomic force microscopy (AFM) showed smoothing and step bunching on (100) and (111) surfaces. Natural diamond powders, which were quite irregular and rough prior to treatment, remained the same size but became markedly smoother and well-faceted, as observed by scanning electron microscopy (SEM). The degree of faceting was sensitive to plasma power level or filament temperature and substrate temperature but was independent of H 2 flow rate. The size and degree of faceting appeared to be the same after plasma treatment for isolated and closely packed particles. We argue that surface diffusion is the dominant smoothing mechanism, but clear evidence of etching was observed at substrate temperatures of 975°C and above.

Role of surface and interface science in chemical vapor deposition diamond technology

Diamond is well known as the hardest material in nature. It also has other unique bulk physical and mechanical properties, such as very high thermal conductivity and broad optical transparency, which enable a number of new applications now that large areas of diamond can be fabricated by the new diamond plasma chemical vapor deposition (CVD) technologies. However, some of the most interesting properties of diamond, including the ability to be grown over large areas by CVD processes, result not from its bulk properties but from its special and unique surface chemistry. The surface chemistry derived properties are as remarkable as the bulk properties, and in the end may enable the development of new applications, technologies, and industries which are at least as important as those based on the bulk properties. Some of these surface properties are extreme chemical inertness, low surface energy, low friction coefficients, negative electron affinity, biological inertness, and high over-voltage electrode behavior. The surface science and some of the interesting ongoing research in these areas are explored and illustrated, and unresolved questions are highlighted.

Elastic properties of polycrystalline cubic boron nitride and diamond by dynamic resonance measurements

Abstract The elastic properties of ceramic materials are important to their performance in severe abrasion and wear applications and also provide a useful, quantitative measure of quality. We have measured the Youngs modulus E , shear modulus G , and Poissons ratio ν of several commercial polycrystalline cubic boron nitride and diamond products, in the form of free-standing disks, using the dynamic resonance method. The latter method is accurate and fast enough for routine quality control. Measurements on polycrystalline c-BN yielded values of E, G , and ν in the ranges of 630–770 GPa, 270–340 GPa, and 0.14–0.18, respectively, depending on the volume fraction of superabrasive, binder phase and microstructure. As a point of comparison, the orientation-averaged values of E, G , and ν for pure, equiaxed, polycrystalline c-BN are calculated as 909 GPa, 405 GPa, and 0.12, respectively. Measured values of E, G , and ν for sintered diamond lay in the ranges of 915–990 GPa, 415–450 GPa, and 0.10–0.11, respectively. The modulus results are compared to selected additional material properties.

The role of methyl radicals and acetylene in [100] vs. [111] diamond growth

Abstract The interaction of mechanistic experiments and detailed models are greatly improving our understanding of the mechanism of diamond growth by chemical vapor deposition. Methyl-radical models typically predict growth rates on (111) planes that are much smaller than experiments, unless contributions from acetylene in nucleating new layers are included. These models predict rather different contributions of methyl radicals and acetylene to growth on (100) vs. (111) planes. On the other hand, other models predict rapid inter-conversion of adsorbed hydrocarbons and surface migration, and equivalence of the behavior of methyl radicals and acetylene (apart from a sticking coefficient) might be expected. We have nucleated and grown μm-sized diamond particles at 800°C in a flow-tube apparatus that permits growth from only methyl radicals or acetylene in atomic hydrogen, in contrast to the complex mixture of species found in a normal reactor. Growth from methyl radicals only produced cubo-octahedral crystals with an α value (√3× the ratio of growth rates in the [100] and [111] directions) near 1.8, indicating that the absence of acetylene is not a significant impediment in nucleating new (111) planes. Diamond growth from pure acetylene produced octahedra (α=3), indicating that (100) growth is much more facile than (111) growth in the absence of methyl radicals, and the (111) facets had a high concentration of contact twins. The implications of these results for the mechanism of diamond growth are discussed.

CF3 radicals as growth precursors and halogen-assisted growth on diamond (100)2 × 1: an ab initio study

In this paper the first ab initio results for chemisorption of CF3 radicals, their decomposition, and possible initial steps for growth on the diamond (100)2 × 1 surface are presented. The dimer reconstructed surface is modelled with clusters comprising nine and ten carbon atoms in four separate layers, which form the basic structural unit of the diamond lattice. The RCF3 and RCF2F bond energies and heats of reaction for incorporation of the CF2 adspecies into a new unreconstructed diamond layer have been calculated along with the effects of the lattice constraints on surface structures and energetics. Heats of some plausible fluorine transfer reactions have also been determined. The implications of these results for diamond growth by chemical vapour deposition or atomic layer epitaxy from halogen-containing precursors are discussed.

Self-limiting diamond growth from alternating CFx and H fluxes

Abstract Self-limiting deposition (also known as atomic layer epitaxy) of diamond could provide a means for deposition of uniformly-thick films on non-planar substrates at moderate substrate temperatures, with potential applications ranging from cutting tools to microelectronics. We have deposited diamond films on Mo substrates using alternating fluxes of CFx radicals and atomic hydrogen in a quartz-tube hot-filament reactor, observing deposition kinetics that are suggestive of atomic layer epitaxy in certain respects. The CFx radicals, most likely CF3, were generated by reaction of filament-generated atomic hydrogen with an excess of CF4. Under suitable reaction conditions, the growth rate was independent of CF4 duty cycle (CF4/H2 flow ratio) and equalled roughly one monolayer per cycle, indicating that the surface reached a self-limiting state. However, the growth rate showed a residual temperature dependence and displayed a sublinear dependence on cycle frequency, indicating that the nature of the self-limiting state (e.g. surface F coverage) is temperature- and time-scale-dependent. Implications of the present results are discussed along with a proposed mechanism.

Growth of diamond films from a continuous or interrupted CF4 supply

Abstract Self-limiting film deposition chemistry (also known as atomic layer epitaxy) of diamond, proposed but never demonstrated, may provide a means for deposition of uniformly thick films on nonplanar substrates at reduced substrate temperatures and may also be useful for fabrication of sharply delineated doped layers during epitaxial growth, for example. We have deposited diamond films on Mo substrates using continuous and interrupted flows of CF4 in hydrogen in a quartz-tube hot-filament reactor. High quality diamond growth was obtained at substrate temperatures between 725 and 850 °C, CF4 concentrations in the range of 3–30%, and pressures between 5 and 45 Torr. Under certain growth conditions, distinct morphological differences were observed in diamond films grown using continuous versus interrupted CF4 gas supplies. Implications of the present results for self-limiting diamond growth are discussed.

Growth and characterization of blue and near-ultraviolet light-emitting diodes on bulk GaN

Blue and near-ultraviolet (UV) InGaN/GaN multiple-quantum-well light-emitting diodes (LEDs) were grown on GaN and sapphire substrates using metalorganic chemical vapor deposition. The homoepitaxial LEDs exhibited greatly improved microstructural and electrical properties compared to the devices grown on sapphire. As a result of defect reduction, the reverse-bias leakage current was reduced by more than six orders of magnitude. At forward bias, thermally activated current rather than carrier tunneling was dominant in the LEDs on GaN. The improvement of optical characteristics was found to be a strong function of In content in the active region. At low and intermediate injection levels, the internal quantum efficiency of the UV LED on GaN was much higher compared to that on sapphire, whereas the performance of the blue LEDs was found to be comparable. At high injection currents, both the blue and UV LEDs on GaN greatly outperformed their counterparts on sapphire. The homoepitaxial LEDs with a vertical geometry had a much smaller series resistance and were capable of operating at 600 A/cm2 in cw mode due to uniform current spreading and efficient heat dissipation.

Interaction of hydrogen and CFy radicals on the diamond (100)2 × 1 surface: an ab initio study

Abstract In this work mechanistic aspects of the ALE and CVD growth of diamond from CF y radicals are examined by means of ab initio calculations on the dimer-reconstructed C(100)2 × 1 surface. As models for the 2 × 1 reconstructed and unreconstructed (100)1 × 1 surfaces we have used C 9 and C 10 clusters comprising the basic structural unit of the diamond lattice with one surface dimer. The surface dimer and RCF y bond lengths, RCF y and RCF 2 F bond energies, activation energies for abstraction of fluorine from the chemisorbed CF 3 group, and heats of reaction for addition of a carbon adspecies into a new diamond layer have been determined along with the effects of the lattice constraints on surface structures. Heats of a 1,2 fluorine transfer reaction on the surface has also been determined. The implications of the present results for halogen-assisted diamond growth by ALE and CVD are discussed.