: Progress continued in 1992 in the two major thrust areas of the diamond program; diamond consolidation, and heteroepitaxial nucleation. We have been developing a consolidation technology as one approach to large area diamond single crystals. During this phase of the program, techniques were developed for (a) bonding of natural diamond seed crystals to non-diamond substrates, (b) overgrowth for consolidation, (c) ion-implantation for lift-off of thin diamond platelets. Experimental and theoretical progress was also made on heteroepitaxial growth of diamond. Theoretical modeling indicated a possible role of subsurface carbon in Ni(111) in preventing diamond nucleation. Experimental studies have provided evidence for several roles for oxygen on diamond surfaces during diamond growth. Progress was also made in developing techniques for microstructural characterization of CVD diamond. TEM sample preparation methods were developed for cross-sectional and plan-view analysis of diamond films and substrates. A simple method for rapidly assessing defect densities in CVD and natural diamond was also developed.

Ultraviolet and X-ray photoelectron spectroscopy techniques have been employed in a preliminary study of the electronic structure of polycrystalline diamond films that have been grown on Si substrates by rf-plasma enhanced chemical vapor deposition using water/ethanol growth chemistries. In particular, polycrystalline diamond films with distinctly different surface morphologies and Raman scattering characteristics have been investigated. Corresponding ultraviolet photoemission spectra from air-exposed samples have shown the presence of a prominent low-energy secondary electron emission peak indicative of a negative electron affinity (NEA) surface. Chemical stability of the polycrystalline diamond NEA surface has been demonstrated following conventional acid cleans and hydrogen plasma processing. In contrast, an oxygen (20%)/Ar plasma exposure has been shown to extinguish the photoemission of low-energy secondary electrons and remove the NEA. However, by employing a high-temperature anneal at 750 °C for 15 min in ultra-high vacuum the NEA surface can be restored. Compared to NEA single crystal diamond surfaces the photoexcited low-energy electron emission from chemical vapor deposited polycrystalline diamond films is more robust.

Secondary electron (SE) yield was enhanced by a factor of ∼30 and surface conductance increased up to 10 orders of magnitude when O-terminated or non-terminated natural diamond (100) surfaces were exposed to atomic H. The SE yield from atomic H-exposed surfaces was spatially dependent on near-surface microcrystalline perfection enabling defect-contrast imaging in the conventional SE mode of the scanning electron microscope (SEM). Ultraviolet photoelectron spectroscopy (UPS) on atomic H-exposed surfaces revealed an intense low energy peak attributed to photoexcitation of secondary electrons into unoccupied hydrogen-induced states near the conduction band edge and their subsequent escape into vacuum. The low energy photoemission peak, enhanced SE yield and enhanced surface conductivity were completely removed via high temperature annealing or exposure to atomic O creating the denuded or O-terminated surfaces, respectively.

We have continued to investigate different diamond single crystal substrates to assess their suitability for the growth of homoepitaxial diamond. Surface topography and defect microstructure of the resulting films are two important qualifiers for the choice of a suitable substrate. Previously developed methods to prepare plan-view TEM samples from bulk diamond substrates and homoepitaxial films have been employed. This contribution describes some of our most recent observations of the surfaces andmicrostructure of natural and synthetic diamond single crystals and homoepitaxial diamond thin films using SEM and TEM - including spatially resolved cathodoluminescence (CL) results.Figs. 1 and 2 show TEM micrographs taken from a (100) natural type Ila diamond crystal and a (100) synthetic type lb diamond that have been ion milled to electron transparency. The mottled background contrast observed in both images which has been observed previously and is believed to result from the extended ion milling used to produce these thin foils.

An apparent activation energy for CO desorption from (100) diamond surfaces exposed to atomic oxygen was determined by thermal desorption spectroscopy performed in ultrahigh vacuum and found to be equal to 45.0 kcal/mol. A minimum potential-energy reaction path was identified by semiempirical quantum chemical calculations. Starting with an O-on-top radical site, the reaction proceeds through a β-scission of the C—CO bond, formation of a dimer C—C bond, and finally cleavage of the second C—CO bond. The largest barrier along this pathway is that of the final desorption step; it is equal to 38.4 kcal/mol, in reasonable agreement with the experimental activation energy. Taken together, the broad experimental desorption-peak feature and the multitude of possible desorption sites with differing potential-energy barriers, suggests the existence of a distribution of CO sites on diamond surfaces.

Semiconducting diamond films have the potential for use as a material in which to build active electronic devices capable of operating at high temperatures or in high radiation environments. Ultimately, it is preferable to use low-defect-density single crystal diamond for device fabrication. We have previously investigated polycrystalline diamond films with transmission electron microscopy (TEM) and scanning electron microscopy (SEM)e.g.1 and homoepitaxial films with SEM-based techniquese.g.2. This contribution describes some of our most recent observations of the microstructure of natural diamond single crystals and homoepitaxial diamond thin films using TEM.We are developing techniques to thin diamonds to electron transparency in the region of interest; particularly as diamond is difficult to mechanically thin. One simple (but time consuming) approach for the fabrication of plan-view diamond TEM samples is to thin only by ion milling. Figs. 1 and 2 show TEM micrographs taken from a (100) natural type IIb (semiconducting) diamond crystal and a (100) natural type la diamond that have been ion milled to electron transparency.

Diamond films have been deposited in a low pressure, radio frequency (r.f.) induction plasma-assisted chemical vapor deposition system. The r.f.-induction system confines the plasma at the low pressures of operation 0.010–10.00 Torr to permit efficient dissociation of the reactant gases. A variety of chemical systems have been used to deposit diamond, including traditional H2-CH4 discharges containing 0.5–2.0% CH4; H2-CF4 discharges containing 4–16% CF4, and water vapor discharges containing high concentrations of alcohol and/or acetic acid vapors. No molecular hydrogen is admitted to the growth chamber for the water-based processes. The water vapor becomes the functional equivalent of the molecular hydrogen used in more traditional H2-CH4 discharges. The success of the low pressure r.f.-induction plasma for diamond growth from the wide variety of chemical systems is predicated on the generation of a high electron density plasma. Parent gaseous molecules are converted into appropriate high temperature stable products such as H, H2, CO, and C2H2 as they traverse the plasma. Quadrupole mass spectroscopy has been used to study the conversion of the water-alcohol vapors to H2, CO, and C2H2 as they pass through the r.f. plasma, 99% of the CH4O is converted into H2, H2O, and C2H2. These studies show plasma conversion of H2O into molecular H2. The excess oxygen is rapidly converted into CO through interactions of the O, presumably with solid carbon sources. Optical emission from both the water-based discharges and the molecular hydrogen-based discharges shows the propensity for atomic hydrogen generation from these low pressure r.f.-induction discharges.

A hydrogen plasma-based technique for carbon removal has been combined with a modest anneal for oxide desorption at 720 °C to produce atomically clean Si(100)2×1 surfaces. Carbon and oxygen contamination can be removed from silicon surfaces by a 30 s hydrogen plasma exposure at 480 °C (for carbon removal) followed by a 5 min anneal in molecular hydrogen at 720 °C (for oxygen removal). Surface hydrocarbon removal is thought to occur by volatilization through hydrogenation. The mechanism for oxygen removal is believed to be more straightforward and consist of thermal desorption of SiO at approximately 700 °C. Hydrogen plasma exposures on the order of 5 min are seen to induce microscopic surface roughness without complete oxygen elimination. Samples which are devoid of oxygen following the 720 °C anneal are found to reoxidize upon re-exposure to the hydrogen plasma. The origin of the oxygenating species is unclear, but likely sources include, SiO or OH from tube wall erosion, or contamination produced by interactions of atomic hydrogen with the chamber walls. A subsequent anneal at 720 °C effectively removes any surface reoxidation. Anneals at 720 °C with carbon present on the surface results in a surface reaction which complexes the carbon and oxygen on the surface, rendering the surface contamination resistant, under nominal conditions, to subsequent plasma processing or annealing.

A low pressure chemical vapor deposition technique using water-alcohol vapors has been developed for the deposition of polycrystalline diamond films and homoepitaxial diamond films. The technique uses a low pressure (0.50 - 1.00 Torr) rf-induction plasma to effectively dissociate the water vapor into atomic hydrogen and OH. Alcohol vapors admitted into the chamber with the water vapor provide the carbon balance to produce diamond growth. At 1.00 Torr, high quality diamond growth occurs with a gas phase concentration of water approximately equal to 47% for methanol, 66% for ethanol, and 83% for isopropanol. A reduction in the critical power necessary to magnetically couple to the plasma gas is achieved through the addition of acetic acid to the water:alcohol solution. The lower input power allows lower temperature diamond growth. Currently, diamond depositions using water:methanol:acetic-acid are occurring as low as 300°C with only about 500 W power input to the 50 mm diameter plasma tube.

Semiconducting diamond films have the potential for use as a material in which to build active electronic devices capable of operating at high temperatures or in high radiation environments. A major goal of current device-related diamond research is to achieve a high quality epitaxial film on an inexpensive, readily available, non-native substrate. One step in the process of achieving this goal is understanding the nucleation and growth processes of diamond films on diamond substrates. Electron microscopy has already proven invaluable for assessing polycrystalline diamond films grown on nonnative surfaces.The quality of the grown diamond film depends on several factors, one of which is the quality of the diamond substrate. Substrates commercially available today have often been found to have scratched surfaces resulting from the polishing process (Fig. 1a). Electron beam-induced current (EBIC) imaging shows that electrically active sub-surface defects can be present to a large degree (Fig. 1c). Growth of homoepitaxial diamond films by rf plasma-enhanced chemical vapor deposition (PECVD) has been found to planarize the scratched substrate surface (Fig. 1b).

: The development of water-based diamond growth processes have lead to low-temperature, low-power diamond growth using water:acetic-acid:methanol mixtures. These mixtures readily inductively couple allowing low-power (500 W) inductive discharges to sustain diamond growth. Currently, growth at 300 C is possible with these acetic solutions. No serious degradation in diamond quality has been observed as the growth temperatures are reduced from 600 to 300 C. In a parallel effort, surface chemistry studies have addressed the role of atomic O on diamond surfaces. Atomic O readily converts the 2x1 surface states into a 1x1:O terminated surfaces. Oxygen desorbs from diamond as CO at temperatures approximately 300 C colder than atomic H desorption. Desorption of approximately 90% of the oxygen from the surface as CO does not result in surface reconstruction. Only upon vacuum anneal to 1000 C does the diamond surface reconstruct.

Using thermal mass desorption and LEED we have studied interactions of H, C12, and F2 with a silicon (100) surface, and exchange reactions of the gases with adsorbates on the silicon (100) surface. Thermal desorption spectra were measured for surfaces dosed with H, C12, and F2 singly and then for surfaces dosed first with a halogen and then atomic hydrogen. Finally, the reverse sequence was studied, with atomic hydrogen dosing and then the halogen exposure. Results indicate that the molecular halogens C12 and F2 are not effective at removing H from a Si (100) surface. However, for the reverse reaction, atomic hydrogen appears quite effective at removing the halogens.