Search

High crystalline quality heteroepitaxial diamond films were successfully formed by using grid-patterned growth on Ir with dc plasma CVD. The crystalline qualities were evaluated by X-ray diffraction, Raman spectroscopy, and etch pit method. Flat {100} diamond films without unepitaxial crystals were formed through epitaxial lateral overgrowth with grid-patterned nucleation region of 100-μm spacing. Omega-rocking curves using X-ray diffraction showed obvious improvement in tilt and twist components for wider grid-spacing. For 100-μm grid spacing, the lowest tilt and twist were 0.064 and 0.043°, respectively. The average dislocation density of 9 × 106 cm−2 for the diamond film grown with grid spacing of 100 μm was the lowest ever in reported values for heteroepitaxial diamond. A correlation between the distribution of crystallinity and dislocation was discussed from the results of etch pit method, Raman mapping and TEM observation. Finally, the crystallinity was compared with single crystal diamonds grown by CVD and HPHT. By using grid-patterned nucleation, the quality of heteroepitaxial diamond on Ir has reached in the range of homoepitaxial diamond film.

Key factors in C 1s photoelectron spectroscopy for realistic samples of single crystal diamonds are remarked. Basic equations for angle-dependent photoelectron spectroscopy applied to single crystal diamond samples are described in Appendix A. Carbon 1s photoelectron spectroscopic works so far reported for hydrogen-terminated and oxygen-terminated diamond (001) and (111) samples were reviewed placing special attention on surface C 1s components with reference to the key factors. The results showed diversity in C 1s photoelectron spectra so far reported. We had three specific subjects of the study in C 1s XPS; the first is that we have reconfirmed the phenomenon that surface conductive layers resumed when smooth non-doped CVD C(111)-O samples were annealed in vacuum [Diam.Rela.Mate.18(2009)206]. A single C 1s XPS surface component was found for a smooth C(111)-O sample before the vacuum-anneal, which was attributed to surface carbon atoms in C O H bonding. The second subject is that dependence of C 1s XPS spectra on surface sensitivity has been measured for all the samples with different surface roughness of C(001)-O, C(111)-O, C(001)-H, and C(111)-H. The results were converted to the energy difference between the Fermi-level (Ef) and valence band maximum (Ev) on the probing depth from the surface. All the samples showed downward bending of Ev toward the surface. For the C(001)-H samples, this was a reconfirmation of previous work [Surf.Sci.604(2010)1148]. For the C(001)-H and C(111)-H samples, various degrees of downward band bending toward surface were observed and analyzed with two-dimensional band simulation. It was concluded that another source of holes such as shallow acceptors is present in a deeper region of the surface in addition to holes very close to surface caused by the charge-transfer-doping. The third subject is that C 1s XPS spectra for superconducting C(111)-O samples showed a lattice distortion of ~9 monolayers near the surface.

A detailed study of diamond growth on diamond microneedles was conducted using micro-Raman spectroscopy of the microneedle and coalescence regions to approach the production of freestanding diamond substrate by heteroepitaxy with the microneedle growth method. The high-density non-diamond-phase carbon is contaminated in the initial stage of overgrowth of diamond on diamond microneedles, but this completely disappears through the quick recovery of the crystallinity of the overgrown diamond layers during the coalescence with the lateral direction growth. We also point out the possibility that a strong driving force is applied to the dislocations generated at the regrowth point and at the coalescence front to enhance the mutual annihilation of dislocations. In addition, we reveal that the stress state changes from compressive stress in the initial diamond layers to a nearly stress-free state in the bulk layers on the microneedles through a momentary tensile stress state at the regrowth point at the tip of the microneedle. Overall results indicate the strong feasibility of producing freestanding, stress-free, single-crystal diamond substrate by heteroepitaxy.

In-situ reflectance interferometry was utilized in this study to understand the characteristics of the reflectance profile for the growth of heteroepitaxial diamonds. We observed the time-resolved Fabry-Perot oscillation of the growing heteroepitaxial diamond. Oscillations in the reflectance profile provided information on the grown thickness and provided the growth rate in real-time. The obtained in-situ profile was experimentally compared with the actual growth status via ex-situ observations and the strong correlation between the in-situ profile and the actual growth status was suggested. The analysis of the profiles provided a reasonable prediction to whether there would be any unexpected growth behavior in real-time for heteroepitaxial diamond growth.

Dislocation propagation during epitaxial lateral overgrowth (ELO) of diamond on Ir using stripe nucleation with two different orientations was investigated by cross-sectional Transmission Electron Microscopy and Etch Pit Method. In case of 〈100〉 stripes with diamond {100} growth sector, [001] dislocation and large inclined dislocation were observed. For 〈110〉 stripes with {111} growth sector, dislocation bending was observed with [110] propagation preferred. These results demonstrate that stripe direction in ELO affects dislocation propagation during diamond growth. The control of growth sector can be used to influence the dislocation distribution in heteroepitaxial diamond.

A new concept for producing freestanding diamond substrate by heteroepitaxy is proposed. Thick diamond growth by heteroepitaxy is often prevented by heteroepitaxial-strain-related substrate bowing as it leads to substrate cracking. The possibility of using diamond microneedles as a mechanism to neglect and/or exploit substrate bowing is discussed in this new concept; the self-breaking effect of the microneedles is proposed as an application to prevent the main bulk diamond layers from cracking. With an aim toward the realization of this concept, the present study shows the first two important experimental verifications through a homoepitaxial experiment with a high-pressure high-temperature diamond substrate: (1) fabrication of high-aspect-ratio diamond microneedles (2 and 100 μm in diameter and length, respectively) by a thermo-chemical etching reaction and (2) overgrowth of diamond on the microneedles with air gaps remaining between microneedles after continuous diamond film overgrowth. We also detail a growth scheme that illustrates how continuous diamond films are created from the microneedles. The strong feasibility for applying the concept in actual heteroepitaxy is suggested through the present study.

The classification of etch pits formed by hydrogen plasma etching on heteroepitaxial diamond has been done by cross-sectional transmission electron microscope (TEM). We demonstrated that the origin of etch pit was mainly [001] threading dislocation. From invisibility criterion of dislocation contrast in TEM observation, this dislocation was identified as edge and 45° mixed dislocation. The correlation between dislocation types and etch pit shape was discussed.

Although Ti-based ohmic contacts are often used in the fabrication of diamond electronic devices, researches on the barrier height (ϕB) of this contact are very limited and no consistent values have been reached. In this study, a direct determination of ϕB was performed using X-ray photoelectron spectroscopy. An array of μm-size Ti/Au ohmic electrodes was made on a lightly boron-doped p-type diamond (001) surface, and C 1 s and Au 4f7/2 XPS spectra were measured. The peak binding energies of the spectra from the sample and from reference samples were compared, and ϕB was determined to be ~ 0.63 ± 0.13 eV for the p-type diamond (001). The result is compared with those of previous works and discussed. Prime novelty statement The Ti-based ohmic contacts are often used in the fabrication of diamond electronic devices. Although reasonably low specific contact resistances between 10− 5 and 10− 7 Ωcm2 at room temperature have been obtained for the Ti-based ohmic contacts, the barrier height of the contact has not been determined consistently. In this study, the barrier height of the Ti-based ohmic contact on a p-type diamond (001) is directly determined by X-ray photoelectron spectroscopy to be 0.63 ± 0.13 eV.

The potential of patterned nucleation growth (PNG) technique to control the wafer bowing of free-standing heteroepitaxial diamond films was investigated. The heteroepitaxial diamond (100) films were grown on an Ir(100) substrate via PNG technique with different patterns of nucleation regions (NRs), which were dot-arrays with 8 or 13 mm pitch aligned to < 100 > or < 110 > direction of the Ir(100) substrate. The wafer bows and the local stress distributions of the free-standing films were measured using a confocal micro-Raman spectrometer. For each NR pattern, the stress evolutions within the early stage of diamond growth were also studied together with a scanning electron microscopic observation of the coalescing diamond particles. These investigations revealed that the NR pattern, in terms of pitch and direction of dot-array, strongly affects the compressive stress on the nucleation side of the diamond film and dominantly contributes to the elastic deformation of the free-standing film. This indicates that the PNG technique with an appropriate NR pattern is a promising solution to fabricate free-standing heteroepitaxial diamond films with extremely small bows.

The method of electron-beam-induced current (EBIC) was used to visualize the defect sites on a p-type (boron-doped) diamond (001) film. For this purpose, an Ag-Schottky layer (~ 2 mm × ~ 2 mm × ~ 50 nm) was deposited on the oxygen-terminated p-type diamond (001) film and used as a source of EBIC signal. The signal current of EBIC image appeared to be as large as ~ 1200 times that of the incident electron-beam current and the difference range in image intensity was also large (1–1200). The observed EBIC images showed many kinds of signatures that are possible ‘killer’ defects for Schottky devices. In order to identify ‘killer’ defects in the EBIC image, an array of Ag-dots (~ 40 × ~ 50 μm 2 ) was deposited on an oxygen-terminated p-type diamond (001) film and I–V characteristics were measured on 53 Ag-dots. The resulting I–V characteristics showed that 21 Ag-dots reside on ‘killer’ defects. Comparison between the EBIC image and the positions of Ag-dots residing on ‘killer’ defects showed that large dark dots in EBIC image correspond to the position of ‘killer’ defects. The number density of the large dark dots (i.e., ‘killer’ defects) was ~ 10 4 /cm 2 in the present sample. It is suggested that a high yield Schottky-junction device may be fabricated by avoiding these ‘killer’ defects by the use of EBIC.

Diamond thin film has outstanding physical and chemical properties. Diamond-on-iridium configurations have been prepared by several methods, such as microwave enhanced plasma CVD, direct currency plasma CVD, and hot filament CVD. In this study, an Ir interlayer was deposited on single crystal sapphires (Al 2 O 3 ) with A-planes {1120} by an RF magnetron sputtering method after annealing samples. In addition, a diamond thin film was deposited by a microwave enhanced plasma chemical vapor deposition (MPCVD) method using a mixture of hydrogen and methane gases after a bias enhanced nucleation (BEN) procedure. Ir (001) was grown on the A-plane of sapphire by X-ray pole figure measurement. Diamond thin films were synthesized on each Ir/sapphire substrate and characterized by SEM, Raman spectroscopy. D {100} faces were exhibited in substantial areas of diamond films, and a flat D {100} plane was partially obtained. It is considered that diamond thin films on Ir {100} were mainly grown towards the direction and were epitaxially grown in part.

The effect of bias-treatment (BT) on Ir(001)/MgO(001) substrates in a newly invented planar-diode DC-plasma system has been characterized in-situ and ex-situ by X-ray photoelectron diffraction (XPD), XPS, LEED and SEM. Features of XPD patterns of C 1s core levels were in good agreement with those of three-electrode BT [Diamond Relat. Mater. 13 (2004) 2081], although the degrees of anisotropy of C 1s XPD were smaller. Thicknesses of carbon films estimated from intensity ratios of C 1s/Ir 4d 5/2 (or 4f) XPS peaks were about 2 times larger than those of three-electrode BT. LEED patterns showed no diffraction spots after BT. As a result, we conclude that epitaxial diamond crystallites with the size of a few nm or so are embedded in a non-oriented carbon layer. In the cases where no finite anisotropy of C 1s XPD was observed, no epitaxial diamond grains were grown in post-CVD as revealed by ex-situ SEM. Thus, it is concluded that the anisotropy of C 1s XPD can be a useful measure of diamond nucleation by BT on Ir(001) substrates.

Three elastic stiffness coefficients of nanometer grain-size polycrystalline CVD diamond (NCD) films and micrometer grain-size diamond (MCD) film were measured by resonant-ultrasound spectroscopy coupled with laser-Doppler interferometry (RUS/Laser method). The diagonal stiffness coefficients C 11 and C 44 of the NCD films are found to be smaller than those of bulk polycrystalline diamond and MCD film, and decrease as the film thickness decreases. The off-diagonal coefficient C 12 of the NCD films was much larger than those of the bulk diamond and MCD films. Large C 12 was consistently explained by the micromechanics model with random-oriented pancake-shape graphite plates at the grain boundaries. This indicates that the NCD films consist of sp 3 -bonded diamond grains and sp 2 -bonded grain boundaries. Raman spectra showed two peaks due to the D-band and G-band of sp 2 -bonded graphite phase. These peaks increased with an increase of C 12 . Thus, the coefficient of C 12 can predict the grain boundary chemistry or structure of the CVD polycrystalline diamond film.

A simple method of sample treatment for angle-scan X-ray photoelectron diffraction (XPD) for insulating diamond is demonstrated. The resulting XPD patterns have shown that a quantitative measure of near-surface atomic order is obtained for the insulating diamond. This simple method may be applied in principle to any insulating material, which enables non-destructive characterization of near-surface order of insulating materials by XPD.

A novel microfabrication method for heteroepitaxial diamond was developed. Substrates used in present work were heteroepitaxial Ir(001) films deposited on MgO(001) by sputtering method. The Ir surfaces were treated with ion irradiation of CH 4 /H 2 gas by dc discharge using planar diode. Resist masks were patterned on the ion irradiated Ir surface by electron-beam lithography. The mask patterned Ir surface was etched by Ar ion to remove the irradiated surface at non-patterned area using an ion beam etching system. Then, diamond was grown on the substrate by dc plasma CVD method. As a result, we have successfully fabricated the heteroepitaxial diamond on the patterned areas of the substrate in submicron scale.

The effect of bias treatment (BT) on direct-current plasma CVD diamond growth has been studied in situ by X-ray photoelectron diffraction (XPD) together with LEED and XPS. It was found that C 1s XPD patterns from the sample after BT are similar to those of diamond (001). Coverage of carbon after BT is several tens of ML when BT is very successful. However, LEED shows no diamond (001) spots for the sample after BT. These apparently contradictory findings are explained by the sizes of the diamond (001) crystallites, which, after BT, are large enough to produce C 1s XPD patterns of diamond, but too small to have coherent interference spots in LEED. It is concluded from this and other information that BT in a DC plasma creates hetero-epitaxial diamond crystallites a few nm or less. These diamond crystallites may be related to the atomically abrupt diamond/Ir interfaces of DC plasma CVD-grown samples revealed by TEM [A. Sawabe, H. Fukuda, T. Suzuki, Y. Ikuhara, T. Suzuki, Surf. Sci. 467 (2000) L845].

We succeeded hydrogenated amorphous carbon passivation of diamond surface conductive layer. Hydrogenated amorphous carbon thin films were formed by r.f. plasma chemical vapor deposition (CVD) on diamond surface conductive layer. It is found that there is no major difference in resistance of diamond surface conductive layer between before and after amorphous carbon formation on the grounded electrode. The thermal stability up to 200 °C in air of diamond surface conductive layer was improved by the hydrogenated amorphous carbon formation.

We have developed a novel microfabrication method by combining a microwave plasma treatment with a reactive ion etching (RIE) method. First, the RIE system was used for fabricating whiskers of single crystalline diamond. Then, the whiskers were exposed to microwave plasma of hydrogen gas. As a result, we have successfully fabricated a nano-scale diamond rod of the single crystal, which has the crystalline facets. The sizes of the ‘nano-rods’ were 50–200 nm in thickness and several microns in height. Morphology of the nano-rods could be controlled by changing the crystal orientation of the diamond substrate.

Heavily nitrogen (N)-doped polycrystalline diamond is reported to have excellent electron emission properties although its emission sites are limited. This report describes the fabrication processes, characterisation and electron emission properties of N-doped heteroepitaxial diamond which posses high potential as broad area emitter devices. We have obtained epitaxial diamond films on iridium (Ir) substrates by hot filament CVD using urea [(NH 2 ) 2 CO] as a dopant for N. The Ir substrates are bias treated via direct current plasma CVD (d.c. plasma CVD) method prior to the growth. Surface of the deposited films is smooth and continuous in scanning electron microscopy (SEM) observation. The epitaxial diamond growth is confirmed from reflective high energy electron diffraction (RHEED) pattern. It is found that the electron emission from heteroepitaxial diamond film is dominated by standard Fowler-Nordheim type emission due to the existence of strong electric field in the vacuum with the value of 26 V/μm.

The electron emission properties of N-doped homoepitaxially grown diamond have been measured and discussed in order to clarify the electron emission mechanism by excluding the effects of polycrystallinity. As a result, N-doped homoepitaxially grown diamond exhibits extremely low threshold electron emission even from extremely flat smooth surfaces without grain boundaries. This result strongly implies that the low threshold electron emission from N-doped diamond should be caused by the resistance of the film rather than by high β whatever the underlying mechanism is.

Electron emission from both the hydrogenated and the oxidized boron-doped heteroepitaxial diamond with smooth surface is studied in order to understand the effect of adsorption on electron emission. B-doped heteroepitaxial diamond is grown on iridium (Ir) by direct current plasma chemical vapor deposition (CVD). A relatively smooth surface is obtained. The Cls peaks in X-ray photoelectron spectroscopy (XPS) for the hydrogenated and oxidized diamond surfaces are 284.1 and 285.6 eV, respectively. The threshold voltage of the hydrogenated diamond surface in the electron emission is lower than that of the oxidized diamond surface. The emission barrier height ratio of the hydrogenated diamond surface to the oxidized diamond surface is estimated to be 2.6 from the slopes of Fowler–Nordheim (F–N) plots. The change in the emission properties might be examined by the surface dipole formed on diamond surface from the XPS results and the F–N plots.

Although we had reported the remarkable low threshold emission from polycrystalline diamond heavily doped with nitrogen (N) [Nature 381 (1996) 140], the problems caused by polycrystallinity still remain for understanding the electron emission mechanism. This paper describes the growth of N-doped homoepitaxial diamond film {100}, {111} and {110}, and their electron emission properties. N-doped homoepitaxial diamond is grown on synthetic diamond by hot filament chemical vapor deposition. Urea [(NH2)2CO] is used as a dopant for N. Atomic force microscope (AFM) observations indicate that the relatively smooth surface morphologies are obtained for all the films. The epitaxial growth of all the film is confirmed using reflective high energy electron diffraction (RHEED) patterns. Reflective electron energy loss spectra (REELS) indicate that the very surfaces of {100} and {111} are diamond while {110} is graphite rather than diamond. Raman spectra suggest that the bulk of the obtained films are diamond. The resistivities of the films are found to be much higher than the detection limit of the system. The relatively low threshold emission was observed even from the smooth surface and the threshold voltage is confirmed to depend on the crystal orientation. It is speculated from the film characterizations and the electron emission properties that the low threshold emission is due to high resistance rather than rough surface and/or grain boundaries.

We obtained an epitaxially grown free-standing diamond platelet utilizing epitaxial diamond film formed on a {100} iridium surface using a d.c. plasma CVD process. Iridium was selected as a suitable substrate material for the heteroepitaxy of diamond based on original criteria. Confocal Raman spectroscopy revealed that the diamond platelet contained little or no non-diamond carbon. The obtained diamond platelet is transparent to visible light and cleavable along the 〈110〉 direction on the surface. The angles between the top surface and the cross-sectional surfaces are approximately 55°, almost equal to the theoretical angle of 54.74° between {100} and {111} planes in cubic crystals. Therefore, the cross-sectional surfaces would be {111} planes of a typical facet for single-crystalline diamond. This means that the diamond platelet we have formed has relatively good crystallinity.

sp 3 and sp 2 -rich nitrogen (N)-doped diamond electron emitters are prepared by modifying the substrate temperature. Their electron emission properties have been measured in order to examine the effect of the sp 2 /sp 3 ratio on the electron emission from N-doped diamond. The obtained electron emitters have been characterized by reflective high energy electron diffraction, X-ray diffraction and Raman spectroscopy. The electrical resistivities of the sp 3 -rich diamond is 100 G Ω cm or higher while that of the sp 2 -rich diamond is 100 Ω cm. In electron emission characteristics, the threshold voltage of the sp 3 -rich emitter is found to be lower than that of the sp 2 -rich emitter. From the threshold voltage versus the anode-diamond spacing characteristics, it has been found that most of the anode voltage drops in the bulk of the sp 3 -rich emitter and almost no voltage drop takes place in the sp 2 -rich emitter. The difference in electron emission properties should be explained by the resistivity of the N-doped diamond bulk.

Electron emission from boron (B)- and nitrogen (N)-doped diamond electron emitter with silver (Ag)-loaded conductive epoxy, titanium (Ti)/gold (Au), aluminum (Al) and amorphous carbon contacts are studied. Both Ti/Au and amorphous carbon contacts on B-doped diamond electron emitters lead to lower turn-on voltage compared to other contacts, while the turn-on voltage of N-doped diamond electron emitter is independent on the backcontacts.

An Ir layer was epitaxially grown on a MgO(100) substrate by radio frequency magnetron sputtering, and was used as the substrate for the epitaxial growth of diamond by microwave plasma-assisted chemical vapor deposition (MPCVD) using CH4–H2 systems. The Ir(100) layer was then bias-treated under a wide variety of conditions. Thus, a bias voltage of −150 V, a methane concentration of 3%, and a continuation of 15 min, were determined to be the optimum conditions. In the subsequent diamond growth step, oriented diamond particles, 100–250 nm in size with pyramidal structures, were formed at a population density of (0.15–1.5)×108 cm−2. It is noteworthy that the fraction of oriented particles was nearly unity. Tilted or rotated diamond particles were seldom observed. Substrate erosion and graphitic carbon formation, also often observed in the case of Ni and Pt substrates, were not found during the diamond deposition on the Ir(100) substrate.

n -Type semiconducting diamond thin films were successfully grown by microwave plasma CVD using phosphine (PH 3 ) as a dopant source. The addition of PH 3 in the gas ambient strongly influences the optimal growth condition for diamond. The {111} surface of diamond was found to be the best surface as a substrate for the growth of phosphorous (P) doped diamond. At a methane to hydrogen ratio of 0.15% and a substrate temperature of 950 °C, smooth {111} oriented homoepitaxial diamond thin films are obtained even with the addition of PH 3 up to 1000 ppm to CH 4 . The epitaxial films grown with PH 3 /CH 4 : 600–20 000 ppm have shown clear n -type conduction by Hall effect measurements over a wide temperature range. An activation energy of electron of 0.46 eV was deduced from the temperature dependence of the carrier concentration. For the best (600 ppm) sample, the Hall mobility of 28 cm 2 /V-s has been obtained at a temperature range around 400–600 K.

This letter presents a clear explanation of the electron emission mechanism of the high-resistivity N-doped diamond cathode. Due to the very low barrier to emission of electrons from the N-doped diamond conduction band into vacuum, electrons in the conduction band of diamond can establish an appreciable leakage current at very low anode voltage. When such a current starts to flow, there is a field which is developed across the diamond bulk. This field is observed as an increase in the electric field at the back contact, causing the injected tunneling current increases exponentially. This process leads to the low threshold emission from the high resistivity N-doped diamond cathode.

A nitrogen (N)-doped diamond-based electron emitter has been fabricated by the sintering technique prior to the chemical vapor deposition process in order to improve the uniformity. There are no spatial differences in reflective electron energy loss spectra (REELS) from the diamond-based electron emitter, suggesting that uniform surface conditions are obtained. The uniform electron emission from the obtained electron emitter is confirmed through emission current vs anode voltage characteristics measurements. It seems that the uniformity of the emitter surface results in uniform electron emission from the diamond electron emitter.

The voltage drop of vacuum and boron (B)-doped diamond electron emitter bulk has been discussed in order to understand the electron emission mechanism of B-doped diamond. It is confirmed that the electron emission from B-doped diamond depends on its film thickness and the localized electron pass/channel is formed in the film. From the results of the threshold voltage versus anode-diamond spacing characteristics, it is found that most of the anode voltage is applied in vacuum, and the high electric field near the B-doped diamond surface is required for the electron emission from B-doped diamond regardless of the film thickness.

The fabrication of a high-quality freestanding diamond substrate was successfully demonstrated via heteroepitaxy by introducing diamond micropatterns and microneedles in the early stage of growth. Micropatterns contributed to a marked reduction in the number of dislocations induced by epitaxial lateral overgrowth, and microneedles relaxed heteroepitaxial strain. Raman spectroscopy indicated the absence of nondiamond carbon inclusions in the obtained freestanding substrate. The full width at half maximum of the X-ray rocking curve for diamond (004) reflections was 0.07°, the lowest value for heteroepitaxial diamond that has been reported so far. The results provide novel insights toward realizing large-diameter single-crystalline diamond substrates.

Heteroepitaxial diamond is grown on a conducting substrate and its electron emission properties are investigated by comparing the electron emission properties of N-doped polycrystalline, homoepitaxial and heteroepitaxial diamond.

The H1 center is the dominant electron spin resonance ~ESR! active center in diamond grown by whatever type of chemical vapor deposition ~CVD! technique. The H1 spectrum has been meticulously characterized in polycrystalline diamond films with a random orientation of the grains ~resulting in a so-called powder-pattern ESR spectrum!. It consists of a central line and two weak satellites ~see Fig. 1!. Remarkably, the peak height intensity ratio ~derivative spectra! of one of the satellite signals to the central line IS /IC decreases with an increasing observational ESR frequency f as IS /IC; f , while the distance between the satellites 2nH increases linearly with f. Taking into account the magnitude of 2nH , the origin of the satellites was unambiguously as

Electron spectro-microscopic methods were applied as direct methods of determining the Schottky barrier heights (SBHs) and their spatial distribution for Au- and Ag-Schottky junctions fabricated on an acid-treated oxygen-terminated diamond (001) substrate. Metal layers were formed with two ranges of thickness (3–5 nm for thin layers and 13–100 nm for thick layers) for both Au- and Ag-Schottky junctions. Leading X-ray photoelectron spectroscopy (XPS) core-level peaks of either Au 4f7/2 or Ag 3d5/2 for the metal layers and C 1s for diamond were used as measures of the SBH. For the thick-metal samples, spatially resolved XPS measurements were performed over most of the sample surface. It was found that there is a variation in SBH on the order of 0.1 eV for the large ("high barrier") SBH values and that there are several places where the SBHs were rather small ("low barrier" junction). For the thin-metal samples, less variations in SBH were observed. The average SBH of "high barrier" junctions for the thick-metal samples appeared to be slightly (0.1 eV order) larger than that for the thin-metal samples. XPS images of leading metal core levels tuned for the "high barrier" and "low barrier" SBHs were observed. For the thick-Ag Schottky sample, the resulting Ag 3d5/2 XPS images clearly showed the locations of defective Schottly junctions. It is suggested that the SBHs determined for the thin-metal samples are the average SBHs on the measured surface and that the SBHs determined for the thick-metal samples are the highest SBHs within the measured µm-size metal islands. The presently determined SBHs were compared with previously reported SBHs and reasonable agreement was found. Photoemission electron microscopy (PEEM) images were observed for the thick-Ag Schottky sample and the "low barrier" islands were identified. The methodologies of XPS, XPS imaging, and PEEM used for the thick-metal samples can be applied to any Schottky junction on diamond.

The selective growth of n-type diamond on a (001) diamond substrate is a powerful method to obtain a lateral p–n junction diode, which is a basic component of power devices. We analyzed the detailed structure and growth mechanism of the selective growth of n-type diamond by cross-sectional transmission electron microscopy. We observed clear boundary lines between the substrate and the newly grown diamond by the selective growth. Thus, the observation of the n-type diamond was possible despite the homoepitaxial growth. It was revealed that the n-type diamond was synthesized in both the [111] and [110] directions from the bottom of the step corner. On the other hand, the growth in the [001] direction was not observed owing to the controlled growth conditions.

Epitaxial lateral overgrowth (ELO) of diamond on Ir(001)/MgO(001) substrates was demonstrated by using a patterned nucleation and growth method. Epitaxial nucleation areas of fine line shape aligned with various crystal orientations were prepared on an Ir(001) surface before diamond growth. The growth rate of the diamonds in the lateral direction markedly changed depending on both the crystal orientation and the growth conditions. A lateral/vertical growth rate ratio of approximately 4.9 was obtained. The full widths at half maximum of the diamond Raman peak observed at the laterally grown areas were approximately threefold better than that of the vertically grown areas on the nucleation sites.

A phosphorus (P)-doped n-type heteroepitaxial diamond (001) sample was grown in a DC-plasma chemical-vapor-deposition chamber. The crystalline properties of incorporated phosphorus were examined by cathode luminescence. Hall-effect measurement of the sample was performed for the temperature range of 293–870 K. The results of these measurements showed that the bulk properties of the n-type sample are comparable to those of previously reported homoepitaxial P-doped n-type diamond (001) samples. The work function of the heteroepitaxial sample was measured by photoemission electron microspectroscopy. The Fermi-level positions at the surface and at a Au Schottky junction of the sample were determined by X-ray photoelectron spectroscopy. Knowing the bulk properties and the critical values of the energy band structure, the whole energy band diagrams at the surface and at the Schottky junction of the sample were evaluated by solving the Poisson equation. The resulting surface electronic structure of the sample was qualitatively different from those of the previously reported H-terminated P-doped n-type diamond (111) surface and heavily P-doped H- and O-terminated diamond (111) surfaces. The presence of ∼0.02 ML of Sn impurity is tentatively ascribed to the peculiarity of the surface electronic structure of the heteroepitaxial sample. The Schottky barrier height of ∼4.2 eV determined for the present sample was consistent with those of previously reported n-type diamond (111) and (001) samples. It is noted that the Schottky barrier height may be determined by the Fermi-level pinning states at the junctions irrespective of the type of dopant to diamond.

The coalescence of epitaxial lateral overgrowth (ELO)-diamonds from stripe-patterned diamond nucleation formed on (001) Ir/MgO has been investigated. Scanning electron microscopy showed only continuous macro steps on the coalesced ELO-diamond layers. Cross-sectional Raman mappings for the 1332 cm-1 diamond peak revealed that ELO regions including coalescence boundaries had a lower full width at half maximum value than regions vertically grown on a nucleation layer. Continuously observed strong tensile strains in ELO regions indicated that additional defect formation to relax the lattice was suppressed even after the coalescence. These results showed that the ELO diamond process could be used to obtain high crystalline diamond films by heteroepitaxy.

Bias-treatment (BT) is widely used as an initial seeding process of chemical vapor deposition (CVD) growth of diamonds. In a BT, substrates are negatively biased by an order of hundred volts so that positive ions in the plasma of CVD gas source bombard the substrate to form 'seeds' of diamond growth. Bias-treated 1-inch-diameter () Ir(001)/MgO(001) substrates were used as the substrates of the CVD growth of 1-inch- diamond (001) thick films. XPS (X-ray Photoelectron Spectroscopy) spectra and C 1s and Ir 4f XPD (X-ray Photoelectron Diffraction) patterns were measured for the BT 1-inch- samples. 5 typical portions on the sample were measured; one at the central portion and the other 4 were peripheral portions, 90° apart from each other and ~4 mm from the edge of the sample. The uniformity of the result over the 5 portions was confirmed. The coverage of carbon layers on the sample after BT was evaluated to be ~ 50 ML in the unit of monolayer of Ir(001)1x1 surface. A quantitative measure of the effect of BT, A45° (the degree of variation of XPD pattern as defined as Imax−Imin/Ia, where Imax, Imin and Ia are the maximum, minimum and average intensities of photoelectrons in an azimuthal diffraction pattern at the polar angle of 45°, respectively), was examined for C 1s and Ir 4f core-levels for the 1-inch- BT samples. The A45° values appeared to be ~3% and ~17% for C 1s and Ir 4f, respectively. The results were compared with those for other bias-treated smaller samples reported so far. The A45° value of ~3% for C 1s is found to be very small compared to the corresponding values for the smaller samples reported so far. Thus, the method of BT for the present 1-inch- sample was found to have a large margin of improvement in terms of A45° for C 1s XPD.

Heavily nitrogen (N)-doped polycrystalline diamond was reported to have excellent electron emission properties. One of the obstacles to practical application of this diamond was its localized and uncontrolled electron emission site. Even though we have reported the broad area electron emission from homoepitaxial diamond in our previous work, the difficulty remained to fabricate the injection contact for homoepitaxial diamond because its substrate is an insulator, which is not suitable for practical applications. In this study, we have obtained heavily N-doped heteroepitaxial diamond films on Iridium (Ir), and its electron emission properties are investigated. It is found that the emission threshold voltage of heteroepitaxial diamond (100) is higher than that of other heavily N-doped diamonds and has linear relationship between the spacing as is not observed in metal–insulator–vacuum type emission, which is a model proposed for electron emission from polycrystalline and homoepitaxial diamonds. From the characterization results, obtained film is confirmed to be identical to heavily N-doped polycrystalline and homoepitaxial diamond. The difference in the emission properties, thus, is caused by high electron affinity of diamond (100) and the absence of negative charges at the Ir/diamond interface.

One of the clear visions of microelectronic device engineering has been to replace the thermal cathode by a “cold” emitting cathode. The main advantages of using a cold cathode are that it is smaller and consumes less energy than its thermal counterpart. For a long time the main obstacle to designing such a device was the inability of achieving a broad area low-threshold electron emission from material of interest. This study reports surprisingly broad area low-threshold emission from homoepitaxially grown N-doped chemical vapor deposited (CVD) diamond (111), which was never achieved with polycrystalline diamond. Moreover, there was a strong correlation between the oxygen absorbed site and emission site. Our results suggest a superhard, metastable planar cold cathode with a desired emission site can be obtained by homoepitaxially grown N-doped CVD diamond (111) with selective oxygen absorption.

The relationship between the electron emission properties of a boron-doped chemical vapor deposited polycrystalline diamond thin film and the oxygen coverage on the diamond surface is systematically investigated. The threshold voltage of electron emission becomes higher with increasing oxygen coverage on the surface. The change of emission barrier height estimated from the slopes of Fowler-Nordheim plots and from the coverage of adsorbed oxygen atoms is evaluated. From the results, it was speculated that the emission barrier height was changed by the surface dipole layer that was formed by oxygen atoms on the diamond surface and led to the changes in the electron emission properties.

Electric properties at metal/n-type diamond interfaces are investigated. Homoepitaxial diamond thin films with n-type conduction are grown on {111} diamond substrates by microwave-enhanced plasma chemical vapor deposition. Gold, copper and aluminum, which have a low reactivity for diamond, are deposited on diamond films. Au and Cu contacts are found to give good n-type rectification properties with a small reverse current. However, a large reverse current is observed for Al contacts. This large reverse current may originate from the lowering of the Schottky barrier height at the Al/n-type diamond interfaces.

Heteroepitaxial diamond (100) with relatively smooth surface morphology is grown on iridium thin film by direct current plasma chemical vapor deposition, and its electron emission is discussed by comparing the electron emission of polycrystalline diamond with the various surface morphology. From the results of electron emission measurements, the emission property from heteroepitaxial diamond is independent of the measurement point, while that from polycrystalline diamond strongly depends on the measurement points. In addition, electron emission from heteroepitaxial diamond was observed at lower voltage compared to that from polycrystalline diamond.

We have successfully grown smooth diamond thin films epitaxially on (001) iridium surfaces through a direct-current plasma chemical vapor deposition process with two steps: ion irradiation pretreatment and diamond growth. The epitaxial areas of diamond thin films with a mean thickness of about 1.5 µ m seem to act as optical mirrors. The average roughness (Ra) of the thin film, as measured by atomic force microscopy, is about 1 nm. Confocal Raman spectroscopy was used to investigate the depth profile of the thin film. Raman bands due to nondiamond carbon components were nominal at the diamond/iridium interface or at other depths.

Epitaxial growth of diamond on iridium thin films was performed by direct-current plasma chemical vapor deposition with ion irradiation pretreatment of the substrate. Pyramidal epitaxial diamond particles with a number density of ∼108 cm-2 were grown on the iridium film. The epitaxial relation is written as (100) diamond//(100) iridium and [001] diamond//[001] iridium. Tilting of the epitaxial relation, as occasionally observed for diamond on silicon or beta silicon carbide, is scarcely observed. Erosion,as observed for diamond on nickel substrates, is not observed. The effect of the ion irradiation of the substrate is discussed briefly.

Diamond thin films have been formed by dc plasma chemical vapor deposition from a gaseous mixture of methane and hydrogen (gas ratio (CH4/H2); 2/100, total gas pressure; 2.66×104 Pa) with a growth rate of ∼20 µm/h. During the course of diamond growth, characteristics of the plasma were measured by the Langmuir single probe method and also by emission spectrometry. From the characterization of the plasma, it was found that the degree of ionization was rather low (∼10-7), and the statistical temperature of hydrogen gas was ∼5×103 K. By a thermodynamic investigation at the gas temperature around 5×103 K, it was revealed that the source gases (H2 and CH4) of over 99% are decomposed to neutral atomic hydrogen and atomic carbon.