ICMCTF1999 Session D1-2: Synthesis, Characterization and Applications of Cubic Boron Nitride and Carbon Nitride Materials

Research on carbon nitride was initially inspired by the predicted properties of its cubic from C₃N₄. Its synthesis in a homogeneous form remains to be achieved. However, in the attempt to achieve this a new class of CN_x thin film materials have been found. In many studies CN_x is found to exhibit unique mechanical properties. In particular, it appears to have "hard and elastic" properties under loading conditions which normally lead to significant plastic deformation in other hard films, e.g. terahedral amorphous carbon.

Structural studies on CN_x films show that those with interesting mechanical properties are not completely amorphous. Residual structure in the form of curved graphene sheet fragments are evident. Analysis of bonding using EELS also shows that the hardness cannot be associated with high degree of "diamond-like" bonding. Recent studies also show that well formed carbon nanotube-like and onion-like structures can also be present within the CN_x matrix. Results obtained to date from a number of groups are reviewed. New results which show that the "amorphous" matrix in CN_x can also show interlinked curved graphene plane structure are also presented. The evidence available to date suggests that a new fullerene-like can be obtained in CN_x films deposited under certain conditions

Hydrogenated carbon nitride films were deposited by reactive d.c. magnetron sputtering in mixed N₂/Ar/H₂ discharges. CN_xH_y films were grown onto Si (001) substrates kept at substrate temperatures of either 100 or 350 °C. The total pressure was kept constant at 2,5 mTorr with gas mixtures of Ar/H₂, N₂/H₂ or Ar/N₂(5%)/H₂ and with the H₂ fraction varied from 0 to 20%. As-deposited films were analyzed with respect to their microstructure, chemical structure and mechanical properties by HREM, SEM, RBS, NRA, XPS, FTIR and Raman spectroscopy, and Nanoindentation.

It was found that the growth rate for all films, except those grown in Ar/H₂ discharge at low temperature, decreased with increasing H₂ partial pressure. With more than ~15 - 20 % H₂ in the gas, no net film growth took place, which can be attributed to chemical etching effects on the growth surface. As a result, species like CH₄, C₂H₂, NH₃ and C₂N₂ were observed by mass-spectrometry during growth. The amount of hydrogen in the films, as observed by nuclear reaction analysis, was found to increase with increasing H₂ partial pressure, while the nitrogen concentration decreased slightly, indicating that more carbon bonds become terminated by hydrogen, as also reflected by a narrower C1s XPS peak as less carbon-nitrogen bonds are present. High resolution transmission electron microscopy also shows that films grown in nitrogen at elevated temperature undergoes a transition from a fullerene-like to a more polymeric microstructure as the amount of hydrogen is increased. The mechanical response, as measured by nanoindentation, also changes from being hard and highly elastic, to more moderate values typical for hydrogenated carbon films.

Carbon nitride (CN_x) thin films have been prepared by cathodic arc evaporation of a graphite cathode under simultaneous bombardment of the growing film by a nitrogen ion beam produced by a Kaufmann-type source. Film deposition was performed varying the ratio of arrival rates of nitrogen and carbon ions, Q_r, the energy of the nitrogen ion beam as well as temperature and bias potential of the substrate. For comparision, film deposition was performed without nitrogen ion beam by operating the carbon arc in a nitrogen atmosphere of appropriate pressure. Film composition was analyzed using elastic recoil detection (ERD) analysis.

For each deposition condition the fluences of carbon and nitrogen atoms were calculated from the ion currents. Comparing these fluences to the number of atoms per unit area found by the ERD analysis the sticking coefficients for carbon and nitrogen could be obtained. In addition, the knowledge of the number of atoms per area was used to determine the density of the samples.

Usually, a reduction of the incorporated rates in comparision to the arrival rates was observed both for nitrogen and carbon. A major reason for that is sputtering which, however, cannot explain the full difference quantitatively. Therefore an additional chemical mechanism has to be assumed. Most probably this mechanism consists of the formation of volatile (CN)₂ supported by the dissociation and excitation of neutral nitrogen species contained in the background gas atmosphere by the arc plasma. At increasing Q_r the incorporation of nitrogen is more hampered than that of carbon. This is probably due to the formation and out-diffusion of N₂ molecules.

CN_x films were deposited on Si(100) substrates using DC magnetron sputtering of a high-purity 99.99% graphite target in a pure 99.999% nitrogen at the pressure p in the range from 0.15 to 5 Pa and the constant gas flow of 75 sccm. The target-to-substrate distance was 100 mm. The discharge current on the target I_m was in the range from 0.5 to 3 Å, the negative substrate bias voltage U_b induced by an RF generator operating at a frequency of 13.56 MHz varied from -300 to -1200 V and the substrate temperature T_s was 600 °C.

We prepared amorphous films, typically 1 to 2 µm thick, substoichiometric in nitrogen (N/C up to 0.35) with high hardness (up to 40 GPa), elastic recovery (up to 85%), good adhesion to substrates and good tribological properties.

Discharge characteristics were measured for both mutually coupled discharges (i.e., DC magnetron discharge and RF discharge dominant in a substrate region) and the corresponding L/l_i values, where L is the sheath thickness and l_i is the ion mean free path, were evaluated to provide information on the energies and fluxes of ions bombarding the target and growing films under various deposition conditions. Optical emission spectroscopy was used to study an occurrence of significant atomic and molecular species, such as N, N₂, N⁺, N₂⁺, CN, C and C₂, near a substrate.

Excepting the values of /U_b/ higher than 700 V, for which the N/C atomic ratio in the films decreases at low pressures (mainly due to preferential sputtering of nitrogen from them) and the deposition rate of the films falls very rapidly at higher pressures, a good correlation between the N/CN concentration ratio in the deposition zone and the N/C ratio in the CN_x films was found. It was shown that the films with a high N/C ratio can have a high hardness only in the case that the energy and flux of nitrogen ions, bombarding the growing films during their deposition, are sufficient for an effective incorporation of nitrogen into the films (/U_b/ = 500 to 700 V, J_m = 7.1 mA/cm² and p < 0.5 Pa).

Electrically neutral, laterally homogeneous plasma beams in which the energy E_ion of the ion component can be precisely varied from some 10 eV up to 1 - 2 keV have been shown to be a very suitable means for a controlled deposition of hard compound films /1/. For the deposition of carbon nitride films the source plasma was operated in nitrogen with admixtures of Ar /2/, but also of Ne, the latter components to enable controlled energy input during film growth. The carbon component was added by sputter injection into the source plasma from a large area graphite target. CN-films were grown on Si(100) for operating the plasma beam source with varying N₂ to Ar or Ne ratios at E_ion = 100 eV or 120 eV, respectively. The film composition was determined by SNMS, and N-concentrations around 45 at% were achieved when operating the source with pure N₂. The binding conditions were studied with XPS. From a detailed analysis of the C1s- and N1s-peak structures the fraction with a binding configuration corresponding to a probably ß-C₃N₄ like phase was found to reach 0.58 for N₂/Ar or Ne ratios around 1:2. By separating the contribution of that phase to the total peaks, an N/C-ratio around 1.33 in agreement with a C₃N₄-phase was always found for that fraction of the films. The optical band gap determined by spectroscopic ellipsometry displayed a maximum value slightly below 2 eV which coincides both with maximum film hardness and maximum content of the C₃N₄-phase. Further indication for the formation of a ß-C₃N₄ like phase was supplied by high energy electron diffraction delivering lattice spacings which agree well with corresponding theoretical predictions. Quite interestingly, a strong influence of the N⁺/N₂⁺-ratio in the plasma beam was found which indicates that a hard C₃N₄-phase is preferentially formed by a well defined energy input via atomic nitrogen. } /1/ F.R. Weber, H. Oechsner, Surf. Coat. Technol. 59 (1993), 183 /2/ F.R. Weber, H. Oechsner, Surf. Coat. Technol. 75 (1995), 704 Electrically neutral, laterally homogeneous plasma beams in which the energy E_ion of the ion component can be precisely varied from some 10 eV up to 1 - 2 keV have been shown to be a very suitable means for a controlled deposition of hard compound films /1/. For the deposition of carbon nitride films the source plasma was operated in nitrogen with admixtures of Ar /2/, but also of Ne, the latter components to enable controlled energy input during film growth. The carbon component was added by sputter injection into the source plasma from a large area graphite target. CN-films were grown on Si(100) for operating the plasma beam source with varying N₂ to Ar or Ne ratios at E_ion = 100 eV or 120 eV, respectively. The film composition was determined by SNMS, and N-concentrations around 45 at% were achieved when operating the source with pure N₂. The binding conditions were studied with XPS. From a detailed analysis of the C1s- and N1s-peak structures the fraction with a binding configuration corresponding to a probably ß-C₃N₄ like phase was found to reach 0.58 for N₂/Ar or Ne ratios around 1:2. By separating the contribution of that phase to the total peaks, an N/C-ratio around 1.33 in agreement with a C₃N₄-phase was always found for that fraction of the films. The optical band gap determined by spectroscopic ellipsometry displayed a maximum value slightly below 2 eV which coincides both with maximum film hardness and maximum content of the C₃N₄-phase. Further indication for the formation of a ß-C₃N₄ like phase was supplied by high energy electron diffraction delivering lattice spacings which agree well with corresponding theoretical predictions. Quite interestingly, a strong influence of the N⁺/N₂⁺-ratio in the plasma beam was found which indicates that a hard C₃N₄-phase is preferentially formed by a well defined energy input via atomic nitrogen.

1. F.R. Weber, H. Oechsner, Surf. Coat. Technol. 59 (1993), 183

2. F.R. Weber, H. Oechsner, Surf. Coat. Technol. 75 (1995), 704

The variation in the electrical properties of carbon nitride films deposited by a Penning-type opposed target reactive sputtering technique as a function of nitrogen incorporation was recently described ¹. It was shown that the resistivity and activation energy of the films increase with nitrogen incorporation and this was related to a transition from metallic to semiconducting behavior. In this paper X-ray photoemission spectroscopy (XPS) and ultraviolet photoemission spectroscopy (UPS) are used to study the valence band spectra of the films. XPS studies do not give a good indication of the states responsible for conduction in the films due to cross-section modulation effects ² . On the other hand, UPS is more sensitive to valence band states and gives a better approximation of the actual valence band spectra of the films. The resistivity of carbon nitride films is found to decrease with negative substrate bias. Electron spin resonance (ESR) is used to investigate the paramagnetic centres of these films; their electrical properties are discussed in terms of their bonding structure by examining what paramagnetic species give rise to the observed ESR spectra.

¹ M.A. Monclus, D.C. Cameron, R. Barklie, M. Collins, presented in the 6th conference on Plasma Surface Engineering, Garmisch, Germany (Sept 14-18, 1998).

² F.R. McFeely, S.P. Kowalczyk, L.Ley, R.G. Cavell, R.A. Pollak and D.A. Shirely, Phys. Rev. B, 9 (1974) 5268.