next up previous contents
Next: 2.2 Two-Dimensional Profiles in Up: 2.1 Statistical Moments and Previous: 2.1.2 Probability Weighted Moments

2.1.3 Evaluation of Statistical Moments from Data Sets

  There are many more or less successful methods in microelectronic engineering to measure distributions of dopants or other species. The most commonly used method is the secondary-ion mass spectrometry (SIMS), which allows to measure the one-dimensional depth distribution of a species. One of the tasks for analytical modeling is now to calculate the characteristic parameters for these profiles and manage them in data-bases for further simulations. As analytical ion implantation is still used since the early 1970s, there are parameter sets available in literature for distributions based on central moments [Kle91] [Rys81] [Jah81] [Hof75b]. Most of the established data are not able to capture dose dependencies of the profiles. By increasing the implantation dose effects like channeling or damage are becoming significant. To apply the analytical ion implantation method for high dose experiments, it is necessary to obtain and tabularize dose dependent data.

As the use of L-moments is not established in semiconductor modeling, they need to be calculated from experimental data. In practice, L-moments must usually be estimated from a random sample. Let tex2html_wrap_inline4939 be the sample distribution of the ions, where the value x gives the depth of the located ion, and tex2html_wrap_inline4943 the ordered sample distribution. The L-moment of the tex2html_wrap_inline4945 sample tex2html_wrap_inline4947 is defined to be

  equation531

in particular

     eqnarray546

We used results from crystalline and amorphous Monte Carlo simulations as reference to extract the L-moments for various ion-target combinations [Ins94].



IUE WWW server
Fri Jul 5 17:07:46 MET DST 1996