A. Deneuville

Characterization of heavily B‐doped polycrystalline diamond films using Raman spectroscopy and electron spin resonance

Heavily B‐doped polycrystalline diamond films ([B]≳1019 cm−3) are studied by Raman spectroscopy and electron spin resonance. The formation of an impurity band is accompanied by a Fano‐type interference for the one‐phonon scattering. Bands at 1200 and 500 cm−1 are observed in Raman spectroscopy for concentrations above 1020 cm−3. They are related to maxima in the phonon density of states, and are ascribed to disordered regions or crystalline regions of very small size. The concentration of defects associated with the paramagnetic signal observed around g=2.0030 increases drastically above 1021 B cm−3. The Mott insulator‐metal transition is accompanied by the presence of a new paramagnetic signal (g=2.0007 for 2×1020 B cm−3, g=1.9990 for 1021 B cm−3) ascribed to free holes in the impurity band.

Effect of boron incorporation on the “quality” of MPCVD diamond films

We report the morphology and Raman signals of MPCVD (microwave-plasma-assisted chemical vapour deposition) boron-doped diamond films deposited on Si with a B:C ratio in the gas phase ranging from 20 to 10000 ppm. The crystal shape remains cubooctahedral in the whole range of doping, while the growth rate is reduced as B:C increases. Up to B:C = 400 ppm the full width at half-maximum (FWHM) and a very weak 1500 cm−1 component decrease as the B:C ratio increases. From B:C = 1000 ppm, an asymmetric deformation is observed in the Raman spectra and at the same time the FWHM and the 1500 cm−1 component increase with increasing B:C. For the highest B:C ratio of 10000 ppm additional bands appear around 550 and 1220 cm−1.

Electrical conduction and deep levels in polycrystalline diamond films

We have studied the dark conductivity (field, temperature, and frequency dependence), and the photoconductivity in undoped polycrystalline diamond films. Detailed analysis reveals that either of two alternative models can be invoked to explain all the observed features of the dark conductivity. The first model is a Hill‐type hopping conduction involving the presence of discrete acceptor states located at 0.91 eV above the valence band with a density around 1017 cm−3. The second model involves the presence of a band‐tail of acceptor states extending about 1 eV above the valence band. In this case, variable range hopping conduction dominates at low fields with a density of states at the Fermi level around 5×1015 cm−3 eV−1, while space charge limited currents dominate at high fields. The states controlling the dark conductivity give rise to photoconduction with a threshold around 0.85 eV and a peak at 1.1 eV. The shape of the photoconductivity spectrum suggests that lattice relaxation (with a Franck‐Condon s...

Defects and stress analysis of the Raman spectrum of diamond films

Abstract Diamond films were deposited on silicon substrates at 750 °C, by the hot-filament technique, from a reactive CH 4 (0.1–2%) H 2 mixtures. Two wide Gaussian lines around 1330 and 1500 cm−1 with coupled variations in the whole preparation range appeared in the global Raman spectra. They were attributed to intermediate carbon defects in the diamond crystallites, which might control the confinement length of diamond phonons. Their contributions to the diamond line shift and width for all the samples is calculated and compared with the experimental results. The remaining shift is attributed to the stress (up to 1.2 GPa), while the origin of the remaining widening (large distribution of stress or Raman inactive additional defects) is discussed.

Raman spectra of TiN/AlN superlattices

TiN (4.5 nm)/AlN (3, 6, 22 nm) superlattices deposited by DC magnetron sputtering on MgO(001) at a temperature of 850°C exhibit Raman signals. They indicate N and Ti vacancies (as in thick TiN) in TiN1−x layers (x=3±2%). x is higher for the sample with 3-nm thick AlN layers, which is ascribed to N diffusion from AlN (standing close to the TiN interfaces) to TiN. In comparison to Raman peaks of thick AlN, there are split signals of wurzite AlN phase, and a signal from another phase, which might be defective rocksalt AlN standing close to the TiN interfaces. The Raman signals clearly show interactions between AlN and TiN layers.

Chemical vapor deposition of B‐doped polycrystalline diamond films: Growth rate and incorporation efficiency of dopants

The growth rate and the incorporation efficiency of dopants have been studied in the case of chemical vapor deposition of B‐doped polycrystalline diamond films. The deposition rate is found to decrease with the addition of diborane in the gas phase. This is correlated with a modification of the plasma chemistry as observed by emission spectroscopy (decrease in the H/H2, CH/H, and C2/H ratios with the addition of diborane). The concentration of boron incorporated in the films is observed to vary with the square of the boron concentration in the gas phase.

Reality of doping by boron implantation of CVD polycrystalline diamond from a comparison of Raman and electrical measurements

Abstract The effectiveness of doping in polycrystalline CVD diamond by 10 13 –10 16 cm −2 boron implantation at 77 K followed by annealing at 800 °C has been studied by Raman scattering and I(V, T) measurements. The amorphization threshold is found to be located around a boron doping of 3 × 10 15 cm −2 . Subsequent annealing of samples implanted with boron doses below and above this threshold results respectively in doped semiconducting diamond and graphite. In the high temperature range, the activation energies (between 0.61 and 0.15 eV) are discussed assuming a highly compensated semiconductor behaviour. In part of the low temperature range, hopping conduction occurs. The compensating centres are suggested to originate from native as well as implantation-induced defects.

Effect of boron incorporation on the structure of polycrystalline diamond films

Abstract The X-ray diffraction peaks of undoped and boron-doped polycrystalline diamond films have been recorded up to 8 × 10 20 B-cm −3 . The lattice parameter varies slightly according to the crystallographic direction [111], [220] and [311] of the crystalline perpendicularly to the substrate. For undoped films, it is lower in the [220] direction. In all directions, we measure a large expansion coefficient of the diamond lattice versus the boron incorporation [B]. Microstrains appear independent of [B], while coherent domains have a maximum size around 3 × 10 19 B cm −3 . We suggest contribution of both the different of size of boron and carbon atoms (Vegards law) and of holes in the impurity band of boron (but with a positive “deformation potential”) to describe the lattice expansion from boron incorporation.

IR characterization of diamond films on Si substrates

Abstract IR absorption of diamond thin films prepared by the hot-filament technique on Si substrates is recorded. All spectra exhibit an absorption band at 800 cm −1 assigned to the stretching mode of SiC bond. The equivalent thickness of SiC varies from 30 to 250 A, with a maximum when the substrate is exactly fully covered with diamond, in agreement with a model about the C consumption at the substrate surface. All spectra also exhibit a broad band covering the 600–950 cm −1 range and assigned to the stretching mode of WO bonds. Finally, for high CH 4 partial pressure, two absorption bands appear around 1600 and 2900 cm −1 . They are assigned to aromatic compounds, and to the stretching mode of the CH bond in sp 3 configuration and CH, CH 2 , and CH 3 sites (diamond crystallite surfaces), respectively. The maximum H concentration in these films was determined to be 0.7%.

Raman study of diamond films deposited by MPCVD: effect of the substrate position

Abstract Diamond is deposited from 100 sccm H 2 /0.5 sccm CH 4 or 100 sccm H 2 /4 sccm CH 4 /2 sccm O 2 by microwave plasma assisted chemical vapour deposition, either in the middle of the plasma ball or in a remote position at the level of the bottom wall of the wave guide. When H 2 /CH 4 mixtures are used, from Raman spectroscopy the parasitic phases concentration and the full width at half maximum (FWHM) of the diamond line are seen to decrease from the middle to the bottom position. In the bottom position, the diamond crystallite characteristics are further improved by using the mixture containing oxygen: no parasitic phases can be detected, and the FWHM remains between 3 and 4 cm −1 between 680 °C and 880 °C. The concentration of luminescent centers (from the Raman backgrounds) also decreases for a lower deposition temperature, O 2 in the gas, and a remote substrate position, but has a minor effect on the FWHM. These centers are ascribed to defects in the bulk of the crystallites while the FWHM and non-diamond Raman signal are both ascribed to parasitic phases on the crystallite surfaces. Coherent phenomenological models are proposed for the origins of the decrease of the FWHM and of the concentration of the luminescent centers with the previous deposition conditions.