Hexagonal, cubic, rhombohedral, and wurtzite boron nitride (hBN, cBN, rBN, and wBN, respectively) powders were pressed into a Ta foil for introduction into the ultra-high vacuum apparatus that was used in the core-level photoabsorption measurements. Samples prepared in this fashion were free of possible B or N containing contamination, and were less susceptible to charging during measurement owing to the conductive substrate. The metastable wBN,[9] rBN,[10] and cBN powders that would serve as our spectroscopic standards were measured using x-ray diffraction to determine the phase purity.[11] The cubic material was found to be pure cBN with trace metallic impurities, and the rBN and wBN were found to be 94 and 93% pure respectively, with the remainder being the hexagonal phase source material used in synthesizing these two powders. The crystal structures of these materials are shown in Figure 1.
An incoherent BN/Si film (BN/Si) was prepared using ion-assisted, pulsed-laser deposition according to procedures described previously. [5] The 700Å thick film showed no long-range structural ordering as determined from X-ray diffraction. This indicates that this film had an amorphous, or incoherent structure (i.e., the grain or domain size was smaller than the coherence length of the x-ray probe).[12] Other spectroscopic studies of this fine-grained film using infrared absorption indicate the domain size of the BN films are smaller than the phonon wavelength (about 2000Å) and confirm the amorphous character of this film. Thin-film samples prepared by other methods have produced fine-grained, or amorphous films as well.[13]
The boron and nitrogen 1s photoabsorption from each of these samples were measured using monochomatized synchrotron radiation . These measurements were performed at the IBM/U8 beam line at the National Synchrotron Light Source[14,15] and at the 8-1 and 8-2 beamlines at the Stanford Synchrotron Radiation Laboratory.[16] The NEXAFS from each of the BN samples was measured in one of two ways: a secondary-electron, partial yield mode, and a total-electron yield mode. In both modes, nearly identical results were obtained. Briefly, to measure the core-level photoabsorption cross section the photon energy of monochromatic synchrotron radiation is scanned through the core-level edge while monitoring the electron yield. In the partial-yield mode, the secondary electrons emitted from the samples below 10 eV are measured, and in the total-yield mode, all emitted electrons are recorded. In both instances the low-energy, long mean-free-path electron emission dominates the signal and therefore makes the measurements bulk sensitive and surface contaminant insensitive.
In our earlier report, the boron and nitrogen 1s core-level photoemission was measured and it revealed a B:N ratio of 1:1 in the powders, but was larger for the film. The non-stoiciometry of the thin films is typical of rapid, non-equilibrium growth conditions.[13] The film made for this study was fabricated using a N+ ion assisted growth procedure,[5] and hence, had a B:N ratio close to 1:1.[11] Charging of our insulating BN powders during photoemission reduced the tractable chemical information from this method, however, the NEXAFS measurements are much less sensitive to charging. The B 1s near-edge photoabsorption for the four powders and the BN/Si film is shown in Figure 2. In each spectrum the electron yield was normalized to the energy and temporal dependent photon flux that was measured during the NEXAFS measurement. These spectra were then scaled to the largest feature for comparison purposes. Similar procedures were used to prepare the N 1s photoabsorption data presented in Figure 3. For each measurement, the sample normal was 54o with respect to the synchrotron radiation polarization vector, thereby averaging out any angle-dependence to the absorption intensity.[17]