Because infrared spectroscopy on these BN films has indicated the presence of cBN, it is logical to ask whether the unexplained features in the films' spectra may result from a combination of hBN and cBN states. In fact, the occurrence of a step near 194 ev in the background intensity of the films' B K-edge spectra is probably an indication of the cBN phase. The peak at 199 eV in the BN/Si B K-edge data may also be due to a contribution from cBN. Also attributable to cBN is the increased intensity of the sigma* peak with respect to the pi* peak in the films N K-edge spectra. The size of the above-mentioned deviations from the hBN spectra appears to be correlated with the amount of cBN indicated in each film by the IR measurements.
The three fairly narrow peaks near 192.6, 193.2, and 194.0 eV in the films' B K-edge data (see Table 1) do not have analogs in the cBN or hBN data, nor are there similar features in the K-edge spectra of disordered C films.[4] With much lower energy resolution, Fomichev and Rumsh previously reported a single broad peak near 194.0 eV in their study of hBN powder.[8] The nearly constant energy spacing shows that these peaks are not a Rydberg series[1] and the approximately 0.6 eV magnitude of the spacing rules out an origin involving lattice vibrations. The proximity of these sharp peaks to the pi* excitonic peak suggests a related origin. One possibility is that stacking in the films is disordered and that the individual peaks represent different stacking configurations.
Given that BN films are often non-stoichiometric,[11] another possibility is that these peaks represent a quasi-bound final state formed by a photoelectron and a N vacancy. The most obvious possibility is that these peaks are due to the presence of rBN, whose B K-edge spectrum shows similar small peaks, as documented in Table 1. The unidentified peaks would then be due to modification of the pi bonding by the ABCA stacking that is characteristic of rBN. Positive identification rBN as a minor constituent of these films would require further study.
A closer look at Table 1 and Figures 1a and 2a shows that the peaks for the film synthesized with the higher ion beam voltage and higher deposition rate are considerably broader than those for the other two films. Energetic ions have likely caused damage to this film, corroborating the findings of a previous IR and photoemission spectroscopy study by Wada and Yamashita on ion-assisted evaporation of BN films.[11] Wada and Yamashita also observed metallic B and a smaller amount of cBN in material grown at deposition rates above 1 Å /s.[11] These results are consistent with the observation (see Table 1) that the film synthesized at the higher deposition rate has a lower fraction of sp 3-bonded material. Given the evident degree of disorder in this film, the recent observation of a preferred crystallographic orientation came as a surprise. NEXAFS and transmission electron microscopy experiments showed the hexagonal layer planes were oriented close to orthogonal to the substrate.[7] This orientation is contrary to that usually observed in thin graphite films, where the c-axis is typically normal to the film-substrate interface.[4]