This section starts with a description of the physical processes which are necessary for the generation of an Auger electron.
First of all an (inner-) shell ionisation process has to take place. This is an occurrence with a quite low probability as innershell ionisation cross-sections are small. Following, the ionisation must lead to an emission of an Auger electron. This can either happen immediately or after a radiationless Coster-Kronig transition. For the sake of ease the probability of emitting an Auger electron after ionisation is set to unity in the current version of SESAME. This is not a bad approximation for typical Auger lines in comparison to the error in estimation of ionisation cross sections. Yet it is planned to take these transition probabilities in account for future releases of SESAME.
Auger generation takes place throughout the electron range in the sample, but only Auger electrons generated in the near surface area (some monolayers) have the possibility of leaving the sample without energy losses, and so can be distinguished from the primary electron background. This alone is the reason that Auger spectroscopy is a surface sensitive analytical technique. To avoid simulating Auger electrons generated further inside the sample thus wasting valuable CPU time, only Auger electrons which are generated in a depth from the sample surface not exceeding a certain threshold are taken in account. This threshold is specified in the variable AHISV of the record MODEL in the input deck.
As probability of inner shell ionisations already is small, the probability of the occurrence of an ionisation in the near surface region is extremely small and that it additionally reaches the detector even smaller. To obtain useful results in finite time SESAME uses an approach similar to the single scattering model. For each scattering event SESAME calculates the probability of the generation of an Auger electron. SESAME always creates an Auger electron. The information about this Auger electron is kept in a queue.
When the simulation of the present primary electron is terminated, the Auger electrons held in the queue are simulated. The Auger electrons carry a ''tag'' which carries the probability of their generation. In case an Auger electron reaches the electron detector, only a fraction of it, according to its probability of generation, is counted. This method yields results in short time for all Auger dependent outputs. Yet simulating such an abundance of Auger electrons leads to a much higher usage of CPU time per primary electron, so results dependent on primary electrons (such as energy spectrum of backscattered primary electrons) take significantly longer.
Therefore a variable has been introduced to set the optimal trade-off between simulation of primary electrons and Auger electrons. The variable AUGMUL in the record & MODEL (Chapter 9.5) determines a probability that an Auger electron trajectory is to be simulated for the occurrence of each scattering event. If it is set to one, each scattering event results in a simulation of an Auger electron. If it is set, for instance, to 0.1 then statistically only each tenth scattering event will produce an Auger electron, thus leading to a higher performance simulating primary electrons.
The variable AUGMUL has no direct effect on simulation results, it just influences the fraction of simulation time which is spent for simulating Auger electrons. Setting AUGMUL to 1.0 will lead to the same result as when set to 0.1 but with a statistical error you achieve simulating ten times as much primary electrons with AUGMUL set to 0.1.