LEEM AND LEED SPECTROSCOPY. Both real space images and diffraction images can be obtained as a function of E0 to ob- tain spectroscopic information. These spectroscopic datasets are traditionally labeled LEEM-I(V) and LEED-I(V) respectively, as the Intensity is measured as a function of V0. (Not to be confused with a Current–Voltage measurement in transport physics!) How- ever, for properly calibrated datasets (See Chapter 3), this is a bit of a misnomer, as the electron reflectivity as a function of E0 is measured. In this thesis, such measurements are therefore referred to as spectroscopic LEEM and spectroscopic LEED data. For spectroscopic LEEM data, consisting of real space images, spectra from different areas of the sample can be obtained by measuring the average intensity of such an area for each obtained image. For diffraction patterns, equivalently, spectra of intensity as a function of E0 for different diffracted beams can be measured. Note again the duality between diffraction and real space: in BF/DF-LEEM spectroscopy the diffraction spot selection is done using the contrast aperture during the measurement and the sample area is selected during data analysis, while in µLEED spectroscopy the area is selected using the illumination aperture and the diffraction spot is selected during data analysis. The addition of spectral information to images greatly enhances which information can be extracted from the sample [24, 29]. The energy dependent reflectivity is very sen- sitive to the atomic structure of the sample surface, enabling finger printing of different surface structures to distinguish them. By comparing measured spectra to calculations of the low energy electron reflectivity for different atomic structures, identification of precise atomic structures is possible. Note that the inverse, directly deducing or calculat- ing the atomic structure from LEED spectra is impossible, so to extract a precise atomic structure, the positions of the atoms have to be iteratively refined, calculating reflectivity from every configuration until an optimum is achieved. The computational techniques were originally developed for traditional LEED, which uses slightly higher landing energies. They were therefore optimized by using a simpli- fied model, making such calculations inaccurate for E0 . 80 eV. Increasingly, it has be- come possible to calculate low energy electron reflectivity even for the very low energies occuring in LEEM by using ab-initio scattering calculations, essentially so...
