Fluoreszenzspektroskopische Bestimmung der Flußdichte von zerstäubtem Titan bei reinen und oxidierten Oberflächen

Laser-induced fluorescence (LIF) spectroscopy is customary used for the determination of the flux of ion-sputtered wall material. One of the key problems of this method is, that the fluorescenoe light emitted by the atoms is proportional only to the population density of that electronic state which is excited by the laser radiation. To obtain from the density of groundstate atoms the flux density $\varnothing = n_{1} \vec{v}$ the mean velocity $\vec{v}$ of the atoms has to be determined and to obtain the total flux density, the population of all of the electronic states has to be measured and to be summed up. We haue performed these measurements for titanium sputtered by a 1 keV argon ion beam under perpendicular incidence. lt the tarnet surface is free of adsorbates, the sputtered atoms are neutral and almost exclusively in the groundstate multiplett. This follows from a comparison of the absolute flux density ofthe groundstate obtained from a spectroscopic calibration of the fluorescence signals with the flux of target material obtained from the sputtering of a thin layer of titanium with known thickness. The turn flux measurements agree within $\pm$ 12 %. The systematic error is estimated to be ± 50 This estimate takes the real pulse shape and the spatial and spectral profile of the laser light into account and is supported by a numerical calculation. lf the surface is covered by oxygen, the Ar sputtering yield is reduced by a Factor of 5 - 7. The contribution of the atoms in the groundstate multiplett becomes as small as 8 .5 % of the total density. The low-lying metastable states and the titanium ions contribute up to 33 % and 12 %, respectively. The yield of Ti-Ions is small co€ipared to those values quotedin the literature. This discrepancy is explained by a partial reflection of- the Ti--ions at a positive space Charge potential in front of the target which is produced by the Ar-Iran beam itself due to its high flux density. About 50 % of the total flux of sputtered atoms is detected spectroscopically. We conclude that the other 50 % are ejected as molecules e. g. as TiO. In many cases it is not known if thesurface is clean or oxidized. This might impede the determination of the total flux significantly. However, it is possible by simple criterie to distinguish an oxidized from a clean surface. The population temperature of the finestructure levels of the groundstate e. g. is 1400 K instead of 700 K and alD of Ti I to that of the groundstate increases from 5 x 10$^{-3}$ for a clean surface to nearly 1 for a completelyoxidized surface. In order to supplement these results, we haue extended our investigat.ions to oblique incidence of the Ar-ion beam, todifferent target materials like Fe and Al and to different ions like helium and deuterium. The main conclusion is that in Bach rase only a minority of the sputtered atoms is in the groundstate, if the surface is oxidized. In contrast to this the Ti-atoms sputtered from a nitrogen or hydrogen covered surface are ejected mainly in the groundstate. The widths of the velocity distributions depend an theparticular gas and correlate with the binding energy of the ad sorbate. The principles developed herein haue been applied to thedetection of aluminium in the fusion plasma device Elmo-Bumpy-Torus in Dak Ridge (USA . The velocity distribution of the neutral Al-atoms has been measured and clearly shows that sputtering is the mein release rnechanism. The measured Al-flux-densities near the wall are in accord with those obtained from surface probes