Fix for erroneously non-zero tally results of photon threshold reactions

[nelsonag]

Description

When pushing photons through a model I noticed that threshold reaction tallies (e.g., pair-production) had small, but non-zero, values for energy bins below the threshold.

This can be observed in the attached orig_tallies.out which was produced by running the attached model.xml file.

The cause of the issue is the line where the log of the xs arrays are taken since they are all log-log interpolated. Here, an xt::where call is used to set all zero-values to -500 instead of taking the logarithm. However, when interpolation is applied e is raised to the power of -500, which results in ~7E-218.

If this -500 is replaced with -900, then a correct value of 0 results. This fix is applied here. The result can be
seen in new_tallies.out within the zip.

Checklist

• I have performed a self-review of my own code
• I have run clang-format (version 15) on any C++ source files (if applicable)
• I have followed the style guidelines for Python source files (if applicable)
• I have made corresponding changes to the documentation (if applicable)
• I have added tests that prove my fix is effective or that my feature works (if applicable)

[paulromano]

Thanks for this fix @nelsonag! I made one modification to handle the possibility that the cross section library stores zero photon cross sections as a small positive value (exp(-500)), which can happen for photon libraries derived from ACE files due to #2433. If you concur with the change I made, I'll go ahead and merge.

[nelsonag]

Thanks for this fix @nelsonag! I made one modification to handle the possibility that the cross section library stores zero photon cross sections as a small positive value (exp(-500)), which can happen for photon libraries derived from ACE files due to #2433. If you concur with the change I made, I'll go ahead and merge.

Love it and I concur with the change. Thanks!