An alternative stochastic doppler broadening algorithm

The temperature dependent, ideal gas scattering kernel for heavy nuclei with pronounced resonances was developed, proved and implemented in the data processing code NJOY from which the scattering probability tables were prepared. Those tables were introduced to the well known MCNP code. The analytic solution of this scattering kernel is consistent with the Doppler Broadening method of D. E. Cullen and C. R. Weisbin. In this study we present an alternative stochastic algorithm based on MCNP subroutines which allows for Doppler broadening of the integral as well as double differential cross section. The analytical solution of the kernel is confirmed. This stochastic method is then introduced into the MCNP code by means of a rejection method, suggested originally by W. Rothenstein. The differences between the new kernel and the standard MCNP kernel are illustrated for specific resonances of U-238, Th-232, Au-197 and Hg-199. LWR unit cell calculations are performed and criticality, reaction rates and Doppler reactivity coefficients are calculated. Based on comparison with the analytical (probability S(α,β)-tables) approach it is confirmed that the stochastic algorithm gives accurate results. The differences lie within 1-2 standard deviations for all practical cases that were analyzed. (authors)