Valence-band emission

The problem is that photoemission spectroscopy may meet some difficulty in unraveling the f-p mixing in cases when the broad valence band emission and the 5 f emission overlap. The former, in fact, is described well by the ground state one electron density of states the latter reflects, in general, not the initial state but rather the final state after the excitation. Band as well as cluster calculations yield at present, on the other hand, only ground state properties with sufficient precision. 

Examples of XPS spectra recorded from in situ sputter-deposited ZnO and ZnO Al films are shown in Fig. 4.6. Absence of contaminations is evident from the survey spectra, which show only emissions from Zn and O. Aluminium-doped ZnO films, deposited from a target with a nominal Al concentration of 2 wt%, also show a small Al signal. Detailed spectra of the Zn 2p3/2, O Is, Zn LMM Auger, and valence band emissions for largely different deposition conditions are given in the lower half of Fig. 4.6. 

Constant initial state techniques require the photon energy and electron energy analyser to be scanned synchronously so that hv — E is kept constant. In this way, the photon energy dependent partial yield of electrons in an energy window AE at a fixed initial state energy E = E — hv is measured. Core level to empty surface state transitions are enhanced by selecting the appropriate E corresponding to a minimum in the valence band emission. 

In 1978, the same authors (Platau and Karlsson, 1978) performed a thorough XPS and UPS study of the valence band emission of Ce. Using the different energy dependence of photoemission cross sections of the 4f level and the sd... 

In the work on TiC and related compounds discussed above the structure of valence band emission spectra is not discussed. In several instances, however, it appears that lines originating in the valence bonding region can give information on the order and constitution of bonding molecular orbitals. This has been discussed in some detail by Urchd i) for compounds of second short period elements and by Fischer - for 3d transition element compounds. 

The above discussion also holds for valence band emission, with the exception of the line positions. In angle-resolved photoemission spectra, the peak positions move on the energy scale when the emission direction relative to the crystal axes is varied. From this peak dispersion, the energy versus momentum relation e k) of the solid bulk or surface can be deduced by simple kinematical arguments (Section 3.2.2.4.2). 

In valence band emission, the initial-state wave functions are no longer atomiclike with defined angular momenta. Nevertheless, simple arguments applied to the matrix element can help to characterize their symmetry by means of the photon polarization dependence of photoemission intensities. In such experiments, the electron emission direction is placed within a mirror plane of the crystal surface. The final state f (cf, k) has then, by simple geometrical arguments, even symmetry with respect to the mirror reflection. In both Eqs. (3.2.2.7) and (3.2.2.8), the operator has odd symmetry with respect to the direction of the electric field vector . Therefore, if < is chosen perpendicular to the mirror plane (s polarization), a matrix element of the type mj even odd even) will result in zero transition probability. In this geometry, only initial states j( j) of odd symmetry are seen, while emission from even states is completely suppressed. On the other hand, even states are selected when e also lies in the mirror plane (p polarization). This method can be particularly useful for the identification of molecular orbitals in adsorbed molecules, illustrated in the following example of CO adsorbed on a... 

---