Core excitons

Stizza et al. (73,274) have investigated amorphous vanadium phosphates, which are also of interest in relation to a XAS study of the butane-maleic anhydride (V, P)0 catalysts (99a). From the V K edge useful information is obtained about the distortions in the vanadium coordination sphere [molecular cage effect on the pre-edge intensity (312)] and on the vanadium oxidation state. Notably, V4+ is silent to most spectroscopic methods. A mixed V4+-V5+ valence state can be measured from the energy shift of the sharp core exciton at the absorption threshold of the Is level of vanadium due to Is -f 3d derived molecular orbitals localized within the first coordination shell of vanadium ions. 

Bianconi, A. (1979). Core excitons and inner shell resonances in surface soft x-ray absorption (SSXA) spectra. Surf. Sci. 89, 41-50. 

The quantity Uq is often assumed to be equal to the difference between the ionization potentials of the impurity atom and the perfect crystal atom. The above procedure, introduced by Dow and coworkers, has had remarkable success in elucidating the overall trends of deep impurity levels or deep core excitons. 

Despite the potential, experimental spectra of ELNES and XANES have not been fully utilized in order to monitor the local structural and chemical environment. One of the major reasons is the presence of core-hole effects which leads to a redistribution of the PDOS features [10]. In other words, the presence of this effect has been considered as a bottleneck for the full interpretation of the experimental spectra. For example, O Brien et al. compared their XANES spectra of MgO, o -Al203 and MgAl204 at cation L2,3-edge with theoretical DOS obtained by band calculations, but their unoccupied DOS did not reproduce the experimental spectra [11]. Thus, the origin of the major spectral features was concluded to be the formation of a core exciton, i. e., a bound state of the excited electron due to the presence of a core hole. 

Theoretical spectra of ELNES and XANES of MgO were obtaind by Lindner et al. [25] using the MS-calculation. They couldn t reproduce peak A and attributed it to the formation of a core exciton. In our calculations, however, peak A is well reproduced even by the ground state calculation as shown in Fig. 3. This is not a surprising result, since a core-exciton state in this case is essentially a state at the bottom of the normal 3s band which is slightly... 

The convergence limits arc denoted in Fig. 10 as oo and oo. In contrast, the solid shows less discrete features. The condensed phase features are broader in width and shifted to higher photon energy. Both dominant features are due to spin-orbit splitting of the lowest core excitons, reflecting considerable changes in electronic structure from the gas to the condensed phase. This size dependent evolution of spectral features is shown... 

This assignment is confirmed by model calculations based on the non-structural core exciton theory of Resca et indicating that different... 

If the spectrometer has a large acceptance angle, a and ji contributions are mixed and the spectrum represents the total DOS (density of states) of the unoccupied states (including a and jr MOs) [22]. There is some doubt whether the exact absolute position and intensity distribution of the first three peaks (a, b, c in Table 4/5) have to be attributed to core-exciton features or to the local density of states effect [22]. Based on an intensity analysis of boron and... 

Figure 10.9 Near-edge X-ray absorption fine structure (NEXAFS) for clean (a) and hydrogen-terminated (b) diamond (100). The resonances Spi, Srj,4, and Sh imply unoccupied surface states in the band gap for the respective surfaces. For the clean surface, the symmetry of the surface core excitons could even be determined from...

The unoccupied surface states are predicted to lie between 3.3 and 6.0 eV above the VBM, that is, in the upper half of the fundamental band gap [59]. Again, experimental information is obtained only quahtatively based on NEXAFS [68]. Nevertheless, an absorption resonance 2.2 eV below the excitation energy of the bulk core exciton appears to confirm the existence of unoccupied surface states in the band gap (feature Sh in Figure 10.9b). However, this absorption feature is found... 

Experimental data on unoccupied surface states are again only indirectly obtained via NEXAFS [68]. As for the clean diamond (100) surface, a surface core exciton with an excitation energy 4.8 eV lower than that of the bulk core exciton is observed that can be taken as a qualitative confirmation for the existence of unoccupied surface states within the band gap (S in Figure 10.15a). For C(lll)l x 1 H only the... 

When considering surface states in the band gap one should distinguish occupied (donorlike) and unoccupied (acceptorlike) states. Those of the latter type were not directly accessible experimentally so far, but in fact found in band structure calculations of all the surfaces discussed above. Quahtative confirmation of their existence within the band gap was for the (100) and the (111) surfaces obtained from NEXAFS in form of clear surface core exciton resonances. The unoccupied surface states are not electronically active for p-type material, but are expected to become important for n-type diamond. Occupied surface states in the band gap are found only for the clean diamond (111) surface, but can be removed by hydrogen or oxygen termination. All diamond surfaces are semiconducting. In the case of the clean C(lll)2xl and C(110)lxl surfaces, which show symmetric and unbuckled 7T-bonded rows of surface atoms, many-body effects are responsible for the opening of a surface band gap, which caimot be modeled theoretically on the DFT level. Table 10.2 summarizes the reconstructions and surface state distributions of the diamond surfaces discussed above. 

The photoluminescence (PL) spectrum in Figure 1.7 shows a number of lines related to nitrogen-bound excitons and free excitons. SiC has an indirect bandgap, thus the exciton-related luminescence is often assisted by a phonon. Bound exciton luminescence without phonon assistance can, however, occur because conservation in momentum can be accomplished with the help of the core or the nucleus of the nitrogen atom. That is why the zero phonon lines of the nitrogen atom are seen, denoted and Q , in the spectrum but not the zero phonon line of the free exciton. 

Defects with large central cell correction have very localized wave functions. The larger the correction, the more localized the wave function and the higher the probability of interaction between the core (or central cell) and the electron and/or the exciton bound to the defect. Hence the reason why the line is so much smaller than the Q line in the spectrum, as well as the reason why the phonon replicas to the Q-line, is simply a matter of probability, since the central cell correction is so much larger for a nitrogen defect on a cubic site than a hexagonal site. 

Temperature changes do not appreciably affect ultraviolet optical properties of both metals and insulators, although at low temperatures absorption bands associated with excitons and electron band transitions are usually sharper, and frequencies of peak absorption may shift slightly. In the soft x-ray region, transitions of core electrons buried in the interior of atoms hardly notice temperature changes. 

---