Hi Spencer,
Thanks for asking this question and the previous one, because I have also been wondering about how gamma de-excitation works in FLUKA. I’ve worked with codes like DICEBOX in the past that use relatively sophisticated models for the continuum and (optionally) known excited states. Within the given level scheme that consists of known and artificial levels, conservation of angular momentum is applied consistently in DICEBOX-like codes.
I’ll just report on what I found in the FLUKA documentation, but one of the FLUKA developers should have the last word on this. My comment just adds a little detail to this excellent question.
In Sec. 2.1.1 of the manual, Refs. [Fer96, Fer96a] are explicitly given for the gamma deexcitation models, but I found that Ref. [Fer96b] contains much more information about the topic.
In [Fer96b], the sampling of gamma transitions at energies above the known level scheme is described as follows:
A first sampling is performed on the integrated gamma emission proba-
bilities to choose the character (electric or magnetic) and the multipole order
of the emitted photon, and a second sampling is performed to determine the
emission energy according to the selected multipolarity.
It makes sense to choose the EM multipole first and then see which levels are available as final states. The model for the energy-dependent emission probability is given as [Fer96b] :
with the initial- and final-state densities ρ_i and ρ_f, and the gamma-ray strength function f, respectively. In such a continuous-energy model, the angular momentum conservation should be included in a spin-dependence of the final-state density ρ_f (encoding information like “high-spin states at low excitation energies are unlikely”). However, I could not find any information on the spin dependence ("The assumed level density is the same as in the evaporation part"leads to p49-50 in [Fer96b]). It seems like the density for states of all spins is used for the final states, i.e. the second sampling assumes that the transition with the sampled EM multipole can be performed to any lower-lying excited state (compare to the DICEBOX algorithm which uses energy-, spin-, and even parity-dependent level densities). This violates angular-momentum conservation.
Yet, "When known levels are tabulated, the cascade is forced to pass through them."[Fer96b]. From this, I assume that angular momentum is treated correctly for transitions within the tabulated level scheme.
Have I summarized this correctly? Is [Fer96b] still the reference of choice for this topic?
Udo