Auger spectra of the noble gases. Krypton

The Auger single $3d$ vacancy and satellite spectrum for Kr is calculated. It is shown for the single-vacancy M4,5-NN spectra that the calculations are in good agreement with experiment only if certain levels in Kr Iii are reassigned. The level at 171 997 cm^-1 attributed in Moore's tables to (4p)³²D(4d)¹P₁ is assigned to the missing ³P₁ term, and the ¹P₁ is located at 196 500 cm^-1. These reassignments are consistent with expected configuration-interaction effects. As a result the low-energy doublet in the Kr M4,5-NN spectrum is identified as the M4,5-N₁N₁ transition, in agreement with recent energy calculations. Calculations were made to reproduce the observed satellite spectra (3d)(4p)→(4p)³ and (3d)(4p)→(4s)(4p)², with both a statistical distribution of shakeoff and Auger decay of $3p$ vacancies, as population mechanisms for the $(3d)(4p)$ configuration. It was found that the (3d)(4p)-(4p)³ calculations with an initial $(3d)(4p)$ population distribution from Auger decay of $3p$ vacancies were in excellent agreement with experiment, while the statistically distributed shakeoff hypothesis leads to relatively poor agreement. For the (3d)(4p)→(4s)(4p)² neither calculation is in agreement with experiment, because of the neglect of configuration interaction. The summed calculated intensity appears to agree with the summed measured intensity for an initial $(3d)(4p)$ population produced by Auger decay of $3p$ vacancies. Statistically distributed shakeoff leads to a summed intensity far in excess of the summed measured intensity. We conclude that statistically distributed shakeoff must have a probability significantly less than 5%, in contrast to recent calculations indicating 10% shakeoff.